On behalf of the ICTAI-2025 chairs and organizers, we would like to welcome you to the 37th International IEEE Conference on Tools with Artificial Intelligence (ICTAI-2025). For 37 years ICTAI has served as the first of its kind major international forum where the creation and exchange of ideas related to artificial intelligence are fostered among academia, industry, and government agencies. The conference facilitates the cross-fertilization of these ideas and promotes their transfer into practical tools, for developing intelligent systems and pursuing artificial intelligence applications. ICTAI encompasses all technical aspects of specifying, developing and evaluating the theoretical underpinnings and applied mechanisms of the AI-based components of computer tools such as algorithms, architectures and languages.

This article presents a view-independent synergistic model (formal and statistical) for efficiently recognizing complex human activities from video frames. To reduce the computational cost, the number of video frames is subsampled from 30 to 3 frames/s. SKD, a collaborative set of formal languages (SOMA,KINISIS, andDRASIS), models simple and complex body actions and activities. SOMA language is a frame-based formal language representing body states (poses) extracted from frames. KINISIS is a formal language that uses the body poses extracted from SOMA to determine the consecutive poses (motion) that compose an activity. DRASIS language, finally, a convolution neural net, is used to classify simple activities, and an long short-term memory is used to recognize changes in activity. Experimental results using the SKD model on MSR Daily Activity three-dimensional (3-D) and UTKinect-Action3D datasets have shown that our method is among the top ones.

On behalf of the IC-BIBE-2023 chairs and organizers, I would like to welcome you to the 23rd International IEEE Conference on Bioinformatics & Bioengineering (BIBE-2023). For 23 years BIBE has served as the first and a major international forum to facilitate cross-fertilization of Bioinformatics, Bioengineering and Biomedical ideas and technologies and promote their transfer into practical tools for developing systems and applications in areas such as science, engineering, health and business.

Document understanding is based on the analysis of individual document modalities (i.e., document tables, charts, diagrams and images). Document tables are of particular importance for the readers and convey important information about the contents and results of a study in close synergy with natural language text. Document analysis for the extraction of tabular information involves three steps, namely table detection, recognition and understanding. This paper presents a critical review of table detection methods with application document images. The published works are categorized into (a) rule-based approaches, (b) machine learning approaches and (c) hybrid approaches which leverage a combination of both aforementioned techniques. For each category of methods, a reference diagram showing the key processing tasks and their interaction is presented and discussed. The most representative methodologies of each category are tested and evaluated in terms of a set of features associated main with their level of maturity and performance.

Recent modeling of brain activities encompasses the fusion of different modalities. However, fusing brain modalities requires not only the efficient and compatible representation of the signals but also the benefits associated with it. For instance, the combination of the functional characteristics of EEGs with the structural features of functional magnetic resonance imaging contributes to a better interpretation localization of brain activities. In this paper, we consider the EEG signals as parallel 2D string images from which we extract their visual abstract representations of EEG features. This representation can benefit not only the EEG modeling of the signals but also a future fusion with another modality, like fMRI. In particular, the new methodology, called Bar-LG, provides a reduced discretization of the EEG signals into selected minima/maxima in order to be used in a form of tokens for EEG brain activities of interest. A formal context-free language is used to express and represent the extracted tokens for the selected active brain regions. Then, a Generalized Stochastic Petri-Nets (GSPN) model is used for expressing the functional associations and interactions of these EEG signals as 2D image regions. An illustrative EEG example of epileptic seizure is presented to show the Bar-LG methodology's abstract capabilities.

Lately several studies have shown that Autonomous Intelligent Robotic Wheelchairs (AIRW) have offered valuable assistance to people with mobility issues. The assistance offered by these AIRW to these people is categorized according to the needs of each person. One category of such assistance is the rehabilitation exercises. Nosokoma is an AIRW that can offer a set of simple rehabilitation assistances for elderly and people in need. In particular, it can assist a user to perform simple rehab exercises, like “get up”, “turn around”, and “sit down”. Thus, in this paper, we show a rehab method based on the exercises mentioned above and the way that AIRW could collect the data from these exercises. More specifically, pressure sensors are attached on AIRW robotic arms to properly measure the force applied by the human user on the robotic arms, thus AIRW can provide additional lifting assistance to human user for the successful and safe completion of the sequence of get-up and sit-down tasks.

Proper document analysis is dependent on the accurate processing of each individual modalities [1], [2], [3]. Graphics images are such document modalities that illustrate visually information about the behavior of experimental systems, the trend of the financial market and etc. The automatic processing of information graphics aims to provide summarizations for such datasets with applications to document retrieval, enabling reading for visually impaired people and etc. In this paper we review existing chart analysis methodologies that work towards these goals. More specifically, we categorize the chart analysis methodologies into detection, recognition and understanding of graphics images. Graphics detection includes the localization of areas that potentially include chart figures. The area of graphics recognition focuses on identifying the legends and axis of inputs charts as well as the structural characteristics of their curves. Graphics understanding aims to capture the underlying message of input chart images and provide their summarizations. Additionally, we provide a discussion for each area of research about the characteristics and merits of their presented methodologies. Finally, we propose a comparison scheme that is based on subjective metrics that reflect the needs of modern applications.

Mathematical formulas (MF) constitute a crucial part of technical documents and have drawn the attention of the researcher community recently. In their vast majority, researchers focus on their recognition, retrieval, and structure. Another aspect of the technical mathematical formulas, that has received less attention over the years, is the extraction of their mathematical features and the automated understanding of what they represent. A substantial progress in the field of technical mathematical formulas understanding could contribute significantly to two related areas: the scientific knowledge management through a better digitalized archive of technical documents, as well as the experimental research of mathematics, through a solid modeling of the technical formulas. The work that follows is a comparative evaluation of the published articles on the deep understanding of mathematical formulas of technical documents to date. During this study, the scientific work at hand is presented and the main approaches on the subject are discussed. As the unaltered detailed implementation of every effort is prohibitively demanding, a set of criteria is defined to help evaluate and compare the completeness of each work. The study ends with the evaluation of each work according to the defined criteria and thoughts on the subject of mathematical formulas understanding.

Cardiovascular Disease (CVD) is one of the most detrimental health issues the world experiences each year. Approximately 647,000 people die from CVD each year in America according to the Centers for Disease Control and Prevention. In other words, one person every 37 seconds. Globally, 17.9 million people die from CVD each year as reported by the World Health Organization. Financially, America spends approximately $219 billion each year. The American Heart Association estimates the financial cost to reach $1.1 trillion by 2035. These metrics show the need to continue research in aiding individuals affected with CVD. In this paper, we use the theory of our previous work, a wearable ultrasound vest to create a real-time near 3D model of the heart, to create the beginnings of a heart state prediction system. That is, we create a synthetic dataset using the ranges of normal and abnormal heart chamber volumes to calculate external surface areas and generate data via statistical bootstrapping. Then, feed this system into a Machine Learning algorithm called a Constrained State Preserved Extreme Learning Machine (CSPELM). Our results show that we can differentiate between abnormal data (Atrial Fibrillation, Chronic Mitral Regurgitation, and Post-Myocardial Infarction) from normal data, where higher percentages for abnormal and low percentages for normal is the goal, with CSPELM predictions of 60.28-88.33% to 33.61%, respectively.

Arthritis is a very common medical issue which describes joint pain or joint disease. The most common type, being Osteoarthritis, is due to the cartilage in a joint wearing away over time which leaves the bones to rub against each other causing pain and discomfort. According to the Centers for Disease Control and Prevention (CDC), more than 50 million adults have been diagnosed with Arthritis in the United States of America which is 1 in 5 adults. This equates to a financial cost of $304 Billion in 2013 between direct medical cost and lost wages. Over a period of time, it has been observed that an uneven distribution of pressure (body weight) across the foot while walking increases wear and tear to knee cartilage and thus leads to Osteoarthritis. Therefore, we propose a novel smart shoe for temporal identification and correction for people with abnormal walking patterns consisting of uneven pressure distributions. Our smart shoes include five pressure sensors, four vibrators, memory storage, a processor, a Bluetooth transceiver, a power supply, and a Smart Software application on a Smartphone which houses a Decision Tree (DT) for walking pattern predictions. Our Machine Learning approach has a 91.68% accuracy and shows promise for assisting people with Arthritis.

The present investigation shows that gender of stimuli, age, and emotional categories affects the ability of adults and adolescent with Autistic Spectrum Conditions (ASC) to decode facial and vocal emotional expressions. A total of 60 subjects participated to the research: 15 ASC and 15 control adolescents aged between 10-14 years; and 15 ASC and 15 control young adults aged between 20-24 years. Their tasks consisted in decoding: a) 24 adults and 24 children contemporary facial emotional expressions of happiness, sadness, anger, fear, surprise, and disgust; and b) 20 adult's vocal emotional expressions of the same abovementioned emotions (except disgust). Significant differences were observed between ASC and typically developed peers. The data suggest that gender, type (voices or faces) of stimuli, and participants' age affect the emotion recognition process making difficult the definition of a common and shared pattern of emotional expression's recognition compliance among autistic and control groups. These results suggest that efficient and effective e-health technologies need to be able to learn and adapt to user individual traits and subjective needs to offer personalized assistance and support.

In order to achieve a deep understanding of the information presented in technical documents, one must first understand the document's individual modalities, such as tables, graphics, diagrams, etc. Accurate detection and recognition of these individual modalities constitute key processing steps to ensure efficient understanding as well. In this paper, we present a methodology for understanding the deeper associations of tabular information and expressing them in stochastic Petri-net graphs. This research is focused on tables that follow strictly the IEEE format rules. We divide this work into three distinct steps 1) table detection, 2) table recognition, 3) table understanding. For the part of detection, we study different machine learning and rule-based methodologies for classifying images as tables and compare the results from their evaluation. During the table recognition step, we extract all the necessary information from the table and recognize associations between them. We then convert the recognized associations into attributed graphs, natural language text, and SPN graphs. Finally, we present the preliminary results from the conversion of tabular information into the other forms of information representation.

Finding the maximum testing accuracy in Machine Learning has been the goal since its conception. From this goal, neural networks have been the primary source of continual improvements in prediction performance. Traditionally, backpropagation has been the primary way of training neural networks and the Levenberg-Marquardt (LM) backpropagation has become the fastest method. Recently, the Extreme Learning Machine was introduced which randomizes weights and biases of hidden layers and uses the Moore-Penrose generalized inverse of a matrix to calculate the output weights and biases, providing competitive results at significantly faster training times. In this study, we continue our work on the Constrained State-Preserved Extreme Learning Machine (CSPELM) with a Forest optimization (CSPELMF) and ε constraint Rangefinder (CSPELMR). Furthermore, we provide hyper-parameter settings for the CSPELM to optimize accuracy over training time. Our results show that our methods outperformed the LM backpropagation in a majority of the 13 tested datasets and that the CSPELMF and CSPELMR matched or outperformed the CSPELM in all classification datasets.

Technical Documents holistic understanding was always an old challenging problem for computer scientists. The last two decades, the processing of mathematical formulas has received more attention and a number of research efforts have focused on their syntactic analysis and recognition. This paper focuses on the methodologies used to process mathematical formulas in documents, specifically on the parsing methods. Initially, we offer a short overview of the parsing methods selected and classified into two categories (Top-down and Bottom-up). Then, these methods are individually evaluated using a set of features associated with the efficiency, originality and complexity, etc. by introducing an overall maturity score of each methodology and giving us a general idea about the progress in the field.

Cardiovascular health is among one of the most important areas of the medical field today. Approximately 2,303 Americans die from cardiovascular disease (CVD) daily. By 2035, approximately 130 million Americans will be affected by CVD leading to an estimated $1.1 trillion in medical cost. The medical industry is steadily moving toward individual medicine. To provide assistance in the cardiovascular field, this paper presents a methodology for creating a real time near 3D model of the heart while accounting for the thoracic cage obstruction. That is, creating a partial 3D heart consisting of the apical and parasternal views that are viewable in real time on a monitor. We discuss a wearable vest consisting of eight ultrasound transducers in a 4x2 arrangement covering the parasternal and apical views of the human heart. The calculations for the overlap of each view and between both views are provided. Also, generalized formulas for the sequential activation of multiple ultrasound transducers with no overlap are given. The presented methodology is capable of real time near 3D modeling without constant attention from a medical professional where the final goal will be abnormality detection.

In this paper, a syntactic approach for the Electroencephalography (EEG) signal modeling and analysis is presented. EEG is a noninvasive neuroimaging method which directly measures the brain activity in a form of time series of electrical potential oscillations. The assessment of brain dynamics and activity patterns is of crucial importance for the understanding of brain functionality. Over the years, the progress in data mining and processing techniques have contributed to the exploration of the time and frequency domain characteristics embodied in the EEG recordings. Here, a synergistic methodology that enables the analysis of the structural features of EEG and moreover, aims to combine the EEG channels' spatiotemporal dependencies utilizing the Stochastic Petri Nets model, is presented.

EEG has been extensively used to aid the diagnosis of various brain disorders and also, for the identification of brain activities during cognitive tasks. However, the visual evaluation of EEG recordings is a demanding process, susceptible to error and bias due to the human factor involved. The development of EEG analysis methods coupled with data processing and mining techniques have assisted the feature extraction process from EEG recordings. In this paper, a novel method for classification of EEG signals based on features derived from the EEG morphology is proposed. The classification accuracy, as illustrated through experiment evaluation, shows that the proposed method can achieve adequate results and moreover the extracted features can be used collaboratively with commonly used features from time and time-frequency domain to increase the EEG signal's classification performance.

The papers in this special section focus on neuro-informatics which is considered one of the most attractive research fields for scientists, engineers, practitioners and physicians due to its profound importance in healthcare and in our lives. Human curiosity, the BRAIN project in USA with a very large funding budget, and the exponential evolution of computational informatics and nanotech during the last two decades have inspired and motivated many researchers around the globe to contribute with their research to the brain. The papers are associated with the IEEE BIBE-2017 Conference.

The human cerebellum contains almost 50% of the neurons in the brain, although its volume does not exceed 10% of the total brain volume. The goal of this study is to derive the functional network of the cerebellum during the resting-state and then compare the ensuing group networks between males and females. Toward this direction, a spatially constrained version of the classic spectral clustering algorithm is proposed and then compared against conventional spectral graph theory approaches, such as spectral clustering, and N-cut, on synthetic data as well as on resting-state fMRI data obtained from the Human Connectome Project (HCP). The extracted atlas was combined with the anatomical atlas of the cerebellum resulting in a functional atlas with 46 regions of interest. As a final step, a gender-based network analysis of the cerebellum was performed using the data-driven atlas along with the concept of the minimum spanning trees. The simulation analysis results confirm the dominance of the spatially constrained spectral clustering approach in discriminating activation patterns under noisy conditions. The network analysis results reveal statistically significant differences in the optimal tree organization between males and females. In addition, the dominance of the left VI lobule in both genders supports the results reported in a previous study of ours. To our knowledge, the extracted atlas comprises the first resting-state atlas of the cerebellum based on HCP data.

This paper presents a novel graph-based methodology for development of structural and functional brain graphs. Using data from non-invasive neuroimaging modalities, such as MRI, fMRI and EEG, graphs that represent the brain architecture and functionality, are generated. Graph theory-based analysis has been applied with great success for studying the brain's connectivity, organization and dynamics. Towards this direction the presented Local Global (LG) graph methodology combines the local-regional information to compose a global topological representation.

This paper presents a new multimodal Human-Machine Interaction (HMI) scheme-model for the co-operation of a robotic-nurse (here a robotic wheelchair) and its human user. The HMI model processes vocal commands through a Personalized Isolated Word Recognition System (PIWRS) along with the recognition of Body Pose Angles (BPA) for decision-making in real time. In particular, the HMI scheme is able to recognize: (i) a set of voice commands, (ii) a set of body postures and poses and (iii) calculate the appropriate body angles associated to skeletal data obtained through a set of cameras. Furthermore, the HMI scheme receives specific values provided by pressure sensors, which are being utilized by the user throughout the duration of the tasks to be executed that compose the Active Participation System (APS). All these variables are appropriately combined for the safe control of an Autonomous Intelligent Robotic Wheelchair (AIRW) used by people in need. More specifically, the stand-up, turn-around and sit-down are the procedural steps under study.

Humans are used to express their feelings of selfconfidence/ powerfulness or their distress/sadness through either expansive postures that occupy as much space as possible or closing postures occupying as less space as possible to avoid contact. This conduct suggests that feelings of selfconfidence/ powerfulness or distress/sadness change our body expressions/postures. It can be interesting to assess whether the reverse is also true, i.e. the way we arrange our body at a given moment would affect our feelings. The present research reports an investigation on such argument. To this aim, 50 subjects (25 females) aged between 23 and 31 years were requested to adopt either an expansive (high-powered) or contracted (low-powered) posture for as long as 3 minutes and then asked to bet money in a dice game. The results show that assuming high-power poses favors risk tolerant behaviors and rises feelings of powerfulness. This is not true in the case of low-power postures, which engender a sense of stress, sustained by a significant increase of skin conductance levels. Considerations are made on how to exploit these results for psychotherapy and rehabilitation purposes, as well as, for the implementation of artificial intelligent systems operating as tools for well-being and coaching.

The simultaneous recording and analysis of EEG and fMRI data has received considerable attention and nowadays is considered to be the state-of-the-art multimodal imaging method. Due to the complementary nature of these modalities several new perspectives and insights within the field of human brain mapping have been created. In addition, many fusion methods have been proposed in order to combine the aforementioned modalities. In this paper a brief report of the current knowledge of simultaneous EEG-fMRI data fusion along with a literature review of the methods that combines them are presented. The reviewed methods are then evaluated using a number of related features and a maturity score based on these feature is calculated for each method. The results from this survey lead us to use an EEG-fMRI approach to study the space mapping on visually impaired brain during navigation.

Most of the technical documents are composed by several modalities, like diagrams, tables, formulas, graphics, pictures and natural language text. Each of these modalities and their associations significantly contribute to the overall deep understanding of the technical document and the knowledge represented in it. Here for us all these modalities, except NL text, are considered as "images". Thus, each technical document mainly is composed by NL text sentences and "images". Thus, in this paper we present a methodology where all these modalities can be expressed into the same two modalities (natural languages text sentences and SPN graphs) for better associations and deeper understanding of a technical document. This deeper understanding will come from two different contributions. The first unique contribution will be an enrichment of the NL text part with additional NL text sentences extracted from the "images" of the technical document. The second unique contribution will come from the SPM models of these images that enrich the main diagram by generating a simulator for the system that technical document describes.

This work reports on a series of experiments involving 960 participants (aged between 20-30 years and equally balanced by gender), asked to play the receiver role in a modified version of the Ultimatum Game, where together with information on the offer's fairness (e.g. 40 (fair) vs 10 (unfair) of 100 euros), a photo depicted the trader's appearance (trustworthy vs. untrustworthy) and a text provided his moral description (honest vs. dishonest). Receivers were asked to motivate their decision in connection with the appearance, moral judgment, and fairness of the offer, and report on how these variables affected their emotional feelings. Data analysis shows that, in all conditions containing a fair offer, the trader's appearance plays a significant role in the receivers' decisions in terms of acceptance rate. Moral descriptions play a significant role only in conditions containing an unfair offer. However, when asked to motivate their choices, subjects do not feel the interference of the social appearance, rather they provide more or less equal number of motivations with reference to the amount of offers and moral judgments. As for the emotions driving their decisions, non-converging feelings are observed both at intra and inter group level.

Within this paper, we present a novel method for view-independent simple human activity recognition from video frames. In order to tackle the problem, we are going to reduce the number of frames produce by a video sequence, since we are positive that we can identify activities from sparsely sampled sequence of body poses. Then we will use a cooperative set of formal languages. Named SOMA and KINISIS language respectively. SOMA language represents various information regarding the human body (state) in a frame and will assign to it a unique timestamp. While KINISIS language is a sequence based formal language that is going to use the information extracted from SOMA language and segment simple activities into separate actions and then correctly identify each one of them.

This paper deals with the Automatic Deep Understanding (ADU) of technical documents. Here we present a synergistic collaboration between two different modalities, a natural language text understanding (NLU) method and a diagram-image extraction & modeling (DIM) one for the deep understanding of technical documents. In particular, the NLU extracts the text from the document and determines the associations among the nouns and their interactions, by creating their stochastic Petri-net (SPN) graph model. The DIM extracts the diagrams from the document and produces their graph models. Then we combine (associate) these two models in a synergistic way, which leads to the deeper understanding of the technical document.

An effective change in Human Computer Interaction requires to account of how communication practices are transformed in different contexts, how users sense the interaction with a machine, and an efficient machine sensitivity in interpreting users' communicative signals, and activities. To this aims, the present paper investigates on whether and how positive and negative visual scenes may alter listeners' ability to decode emotional melodies. Emotional tunes were played alone and with, either positive, or negative, or neutral emotional scenes. Afterword, subjects (8 groups, each of 38 subjects, equally balanced by gender) were asked to decode the emotional feeling aroused by melodies ascribing them either emotional valences (positive, negative, I don't know) or emotional labels (happy, sad, fear, anger, another emotion, I don't know). It was found that dimensional emotional features rather than emotional labels strongly affect cognitive judgements of emotional melodies. Musical emotional information is most effectively retained when the task is to assign labels rather than valence values to melodies. In addition, significant misperception effects are observed when happy or positively judged melodies are concurrently played with negative scenes.

This research study here deals with the Stochastic Petri-Nets (SPNs) modeling of an interacting task (getting up from a sitting position) in a human-machine interaction model. The machine is an Intelligent Robotic Wheelchair (IRW) appropriately modified to provide assistance to a human user, who utilizes the IRW. SPNs are recognized as useful modeling tools for analyzing the reliability and performance of systems through time, which is the main reason of choosing them. The main use of this model is to simulate the procedure and eventually used by people with disabilities and the elderly with ultimate goal to improve their daily life. They will get up from a sitting position with the help of the wheelchair, which has robotic arms attached to it. If not for the IRW, these people would need the assistance of another person to get up. Simulated results will be provided from the stage of this study.

The increasing occurrence of chronic conditions among the ageing population and people at risk is one of the major challenges for our society and the high cost for its healthcare systems. Prevention, early detection and efficient management of chronic, long-term conditions contribute radically to the individual wellbeing and the economic sustainability of social and healthcare systems. In response to this need, this paper offers a human-machine interaction (HMI) model using Stochastic Petri Nets (SPNs). This HMI is based on a dialogue model between a virtual medical doctor and a patient for the efficient extraction of non-measurable pathological symptoms. Thus, the goal of such a model is the improvement of life critical situations and long delays (or appointments) for certain categories of people in need, like the elderly or people with disabilities.

We present a novel synergistic methodology for the spatio-temporal analysis of single Electroencephalogram (EEG) trials. This new methodology is based on the novel synergy of Local Global Graph (LG graph) to characterize define the structural features of the EEG topography as a global descriptor for robust comparison of dominant topographies (microstates) and Hidden Markov Models (HMM) to model the topographic sequence in a unique way. In particular, the LG graph descriptor defines similarity and distance measures that can be successfully used for the difficult comparison of the extracted LG graphs in the presence of noise. In addition, hidden states represent periods of stationary distribution of topographies that constitute the equivalent of the microstates in the model. The transitions between the different microstates and the formed syntactic patterns can reveal differences in the processing of the input stimulus between different pathologies. We train the HMM model to learn the transitions between the different microstates and express the syntactic patterns that appear in the single trials in a compact and efficient way. We applied this methodology in single trials consisting of normal subjects and patients with Progressive Mild Cognitive Impairment (PMCI) to discriminate these two groups. The classification results show that this approach is capable to efficiently discriminate between control and Progressive MCI single trials. Results indicate that HMMs provide physiologically meaningful results that can be used in the syntactic analysis of Event Related Potentials.

There is an undisputed need for increased portability and flexibility in diagnostic ultrasound imaging due to the continuously evolving point of care (POC) clinical applications. In order to meet this growing need numerous portable ultrasound systems have been proposed. The design of portable systems is a very challenging task, especially for volumetric imaging system which utilize transducers with hundreds and thousands of elements. This study presents a new low-cost, portable volumetric ultrasound system for ambulatory healthcare monitoring which can serve as an enhanced alternative to the existing portable systems. All hardware components along with vital processes that are performed during the image acquisition, are presented in details. Initially, the design of the proposed 2-D phased array transducer is introduced and its performance is evaluated. Secondly, all hardware components of the system are explicitly presented and finally a software-based methodology for image enhancement that is based in image fusion and a super - resolution technique is described. The imaging quality of the designed transducer is investigated through simulations using FIELD II. Additionally, preliminary simulation results of the applied methodology for 3D images, show significant image quality improvement, especially in the regions where recording from all transducers was possible.

This paper aims at investigating on whether and how semantic information conveyed by gestures supports communication effectiveness. The research hypothesis was operationalized as a word retrieval task. To this aim, 140 subjects (73 males, 67 females) aged between 18 and 35 years were recruited at the University of Salerno (Italy). They underwent a memory task after being requested to watch video-clips explicating 15 every-day words through 4 different presentation modes: a) only audio, b) only gestures, c) audio and articulatory (complex auditory) information, and d) audio, gestures, and articulatory (multimodal) information. It was found that semantic information is most effectively retained when conveyed through the multimodal mode, and that the only gestures outperforms the only audio and complex auditory mode. It was also found that females have higher significant recall ability than males, no matter the experimental condition.

This paper focuses on the combination of StochasticPetri Net (SPN) graphs for event association in the context ofNatural Languages Understanding (NLU). Our general goal isto develop a new NLU methodology. In this paper we presentsome of its components which are: the use of AnaphoraResolution (AR), the extraction of kernel(s) based on thestructure of their parse trees, and the synthesis of SPN graphs.Each extracted kernel is represented with SPN graphs. Herewe define when, why and how to synthesize the produced SPNgraphs for event/action association, while preserving theirtiming and information flow. Finally, we provide examples ofcombined and uncombined SPN graphs of different NL textscreated using our proposed methodology.

Combining information from Electroencephalography (EEG) and Functional Magnetic Resonance Imaging (fMRI) has been a topic of increased interest recently. The main advantage of the EEG is its high temporal resolution, in the scale of milliseconds, while the main advantage of fMRI is the detection of functional activity with good spatial resolution. The advantages of each modality seem to complement each other, providing better insight in the neuronal activity of the brain. The main goal of combining information from both modalities is to increase the spatial and the temporal localization of the underlying neuronal activity captured by each modality. This paper presents a novel technique based on the combination of these two modalities (EEG, fMRI) that allow a better representation and understanding of brain activities in time. EEG is modeled as a sequence of topographies, based on the notion of microstates. Hidden Markov Models (HMMs) were used to model the temporal evolution of the topography of the average Event Related Potential (ERP). For each model the Fisher score of the sequence is calculated by taking the gradient of the trained model parameters. The Fisher score describes how this sequence deviates from the learned HMM. Canonical Partial Least Squares (CPLS) were used to decompose the two datasets and fuse the EEG and fMRI features. In order to test the effectiveness of this method, the results of this methodology were compared with the results of CPLS using the average ERP signal of a single channel. The presented methodology was able to derive components that co-vary between EEG and fMRI and present significant differences between the two tasks.

Segmentation of human bodies in images is a challenging task that can facilitate numerous applications, like scene understanding and activity recognition. In order to cope with the highly dimensional pose space, scene complexity, and various human appearances, the majority of existing works require computationally complex training and template matching processes. We propose a bottom-up methodology for automatic extraction of human bodies from single images, in the case of almost upright poses in cluttered environments. The position, dimensions, and color of the face are used for the localization of the human body, construction of the models for the upper and lower body according to anthropometric constraints, and estimation of the skin color. Different levels of segmentation granularity are combined to extract the pose with highest potential. The segments that belong to the human body arise through the joint estimation of the foreground and background during the body part search phases, which alleviates the need for exact shape matching. The performance of our algorithm is measured using 40 images (43 persons) from the INRIA person dataset and 163 images from the “lab1” dataset, where the measured accuracies are 89.53% and 97.68%, respectively. Qualitative and quantitative experimental results demonstrate that our methodology outperforms state-of-the-art interactive and hybrid top-down/bottom-up approaches.

In this paper we present a novel method for describing the EEG as a sequence of topographies, based on the notion of microstates. We use Hidden Markov Models (HMM) to model the temporal evolution of the topography of the average Event Related Potential (ERP) and we calculate the Fisher score of the sequence by taking the gradient of the trained model parameters given the sequence. In this context, the average Event Related Potential (ERP) is described as a sequence of topographies and the Fisher score describes how this sequence deviates from the learned HMM. This alternative modeling of the ERP is used to fuse EEG information, as expressed by the temporal evolution of the topography, and Functional Magnetic Resonance Imaging (fMRI). We use Canonical Partial Least Squares (CPLS) for the fusion of the Fisher score with fMRI features. In order to test the effectiveness of this method, we compare the results of this methodology with the results of CPLS using the average ERP signal of a single channel. Using this methodology we are able to derive components that co-vary between EEG and fMRI and present significant differences between the two tasks. The results indicate that this descriptor effectively characterizes the temporal evolution of the ERP topography and can be used for fusing EEG and fMRI for the discrimination of the brain activity on different tasks.

In the present paper we have focused our efforts in de-noising/despeckling methods for ultrasound images based on a novel multi -- transducer architecture. Frequency compounding schemes have been implemented and tested in two different images and the results are compared with the state of the art software based methods for ultrasound image despeckling. Signal -- to -- noise ratio, SNR, peak -- signal -- to -- noise ratio, PSNR, mean square error, MSE, and the mutual information measures were used to assess the performance of each method. Furthermore, a general scheme for ultrasound image processing based on the notion of sup resolution technique that combines multiple low resolution, LR, images of the same scene to produce one high resolution, HR, image, used in camera technology, is proposed. The nature of the system's architecture has made possible the use of simple but effective compounding techniques for speckle reduction while at the same time sup resolution approach may produce ultrasound images of better quality and higher resolution. Last but not least, the article provides an insight to the major problems and drawbacks of ultrasound imaging systems and how the proposed architecture is capable of effectively dealing with these problems.

Applying engineering precepts to biological systems has spawn the field of systems biology to investigate a network of interacting components, including the coordination of internal systems of living organisms such as endocrine, nervous, and respiratory with gene and gene product expression, and behavior and environmental factors, and understand how these components together contribute to the disease initiation and progression, biological development, and health. Proceeding from systems biology, systems medicine incorporates complex and dynamic biochemical, physiological, and environmental interactions between all components of disease and health that sustain living organisms. The current special issue includes a selected number of papers presented at the 12th IEEE International Conference on BioInformatics and BioEngineering (BIBE 2012), Nov. 11-13, 2012 under a special session with the same theme, in addition to papers submitted following an open call for papers. The Special Issue presents experiences as well as technological and scientific developments stemming from some flagship projects funded by the EU under the FP7 framework programme aiming to bring together researchers working in the fields of infrastructures and technologies for integrative biomedical research, ICT for predictive and translational medicine and the VPH community at large. A total of 15 papers are included under the following scientific subdomains: 1) mHealth,Wearable Systems and Telemonitoring Services (five papers), 2) Medical Imaging (four papers), and 3) Computational Biology (six papers).

This survey and analysis presents the functional components, performance, and maturity of graph-based methods for natural language processing and natural language understanding and their potential for mature products. Resulting capabilities from the methods surveyed include summarization, text entailment, redundancy reduction, similarity measure, word sense induction and disambiguation, semantic relatedness, labeling (e.g., word sense), and novelty detection. Estimated scores for accuracy, coverage, scalability, and performance are derived from each method. This survey and analysis, with tables and bar graphs, offers a unique abstraction of functional components and levels of maturity from this collection of graph-based methodologies.

The topography of the electrical field does not vary randomly with time but rather displays short periods of stable topographical configurations or spatial patterns, known as microstates. The search of such patterns takes place in the high dimensional electrode space with all the problems that comes with it. In this paper, we present a technique for the extraction, detection and representation of EEG microstates based on the Local Global graph(LG graph). We use the Local Global graph to represent the spatial configuration of the topographic map and use a LG graph matching approach to determine the different microstates. The proposed method was applied on the average, over trials, Event related potential recording presenting a positive peak known as P300. Five dominant microstates were identified using our methodology. During the whole period of the P300 peak there is one active microstate which presents frontal and parietal topography which is expected for the phase locked activations of the P300. The LG graph modeling of the EEG topography provides a flexible descriptor for the EEG topography and can be used for the efficient microstate segmentation of the EEG.

Ultrasound imaging (USI) or Medical Sonography (MS) as it is formally called is widely used in biomedical applications over the last decades. Form the Intensive Care Unit minimally invasive monitoring to the recent point of care testing besides the patient's bed. US imaging outcomes can provide clinicians with a thorough view of the internal parts of the human body at a very low expense. In this paper, we insinuate an alternative approach compared to already existing ones of capturing US images. We propose a wearable ultrasound system composed of an array of ultrasound circular 2D transducers, integrated in a belt, for point of care monitoring of the abdominal region and particularly the liver for critical ill subjects. The configuration of the array and the type of the transducers will entail a system capable of providing clinicians with high resolution 2D imaging of the region of interest (ROI) as well as 3D representation of whole organs without theirs assistance.

Ultrasound imaging (USI) or Medical Sonography (MS) as it is formally called is widely used in biomedical applications over the last decades. Form the Intensive Care Unit (ICU) minimally invasive monitoring to the recent point of care testing besides the patient's bed. US imaging outcomes can provide clinicians with a thorough view of the internal parts of the human body at a very low expense. In this paper, we insinuate an alternative approach compared to already existing ones of capturing US images. We propose a wearable ultrasound system composed of an array of ultrasound circular 2D transducers, integrated in a belt, for point of care monitoring of the abdominal region and particularly the liver for critical ill subjects. The configuration of the array and the type of the transducers will entail a system capable of providing clinicians with high resolution 2D imaging of the region of interest (ROI) as well as 3D representation of whole organs without theirs assistance.

Natural language processing and understanding is an attractive field and many techniques and tools for document processing have been developed. Most of the techniques use either statistical models or graph-based approaches. Here we present the modeling of a methodology based on stochastic Petri-nets (SPN) to explain the transformation of a natural language (NL) sentence into a state machine representation as stated in [16]. In particular, we initially convert NL sentences into graphs using the (Agent - Action - Patient) kernel and then we convert the graphs into SPN graph descriptions in order to efficiently offer a model of semantically represent and understand natural language events of a document. The selection of the SPN graph model is due to its capability for efficiently representing structural and functional knowledge.

Research into eyes-free mobile reading devices has grown in recent years. This research has focused mainly on the image processing required by such a device, with a lower emphasis on the user interaction. In this paper, a model of a voice user interface (VUI) for a mobile reading device is presented. Three field studies with blind participants were conducted to develop and refine the model. A formal grammar is used to describe the VUI, and a stochastic Petri net was developed to model the complete user-device interaction. Evaluation and analysis of the user testing of a prototype led to empirically derived probabilities of grammar token usage for the commands that comprise the VUI.

It is generally accepted that evoked and induced activations represent different aspects of cerebral functions during an Event Related Potentials (ERP) experiment. Independent Component Analysis (ICA) has been successfully applied to event related electroencephalography (EEG) to decompose it into a sum of spatially fixed and temporally independent components that can be attributed to underlying cortical activity. A major problem in the application of ICA is the stability of estimated independent components. In this paper we exploited the split-half approach to assess component stability. We used different measures quantifying both phase and energy aspects of the ERP, in order to distinguish evoked from induced oscillations. We applied these measures to the stable independent components derived from a dataset of progressive Mild Cognitive Impairment (PMCI) and elderly controls. We found reduced energy in the induced theta activity in PMCI subjects, in accordance with previous studies. In addition, PMCI subjects presented lower phase-locking values and diminished late alpha band energy in contrast to controls.

Detecting faces has become a common task in human-computer interaction systems. A robust face detection system should be able to detect faces irrespective of illuminations, shadows, cluttered backgrounds, facial pose, orientation and facial expressions. In this paper we describe a method comprising color constancy based skin detection and Local-Global (LG) graph matching in color images to address these challenges. Illustrative examples are also provided as a proof of concept.

A surveillance system for assisting the elderly in remaining independent in their familiar environment is one of the subjects interest in recent healthcare studies. When mature, it is expected that this system will have the ability to track objects that a resident may lose periodically, detect falls within the home, alert family members or healthcare professionals to abnormal behaviors. This paper addresses the early stages and issues of the development of such a system, the physical characteristics of the system that is being designed, early results, and guidance on the future work that will have to be completed in the future.

Security is an important issue related to the storage and communication of data and information. In data and information security, cryptography and steganography are two of the most common security techniques. On one hand, there is cryptography, which is the secret communication between two parties by message scrambling on the sender’s side and message unscrambling on the receiver’s side so that only the intended receiver gets the secret message. Cryptography and cryptanalysis constitute cryptology (or crypto) while steganography and steganalysis make up steganology (or stegano). This paper presents a comparative survey of various cryptology and steganology techniques by analyzing and comparing different methodologies using a set of predefined parameters. It also offers to the reader valuable knowledge on the state of the art techniques used on cryptanalysis.

It is known that the early detection of chronic diseases significantly increases the life span of the elderly and improves the quality of life in general. Since the brain is the most valuable part of the human body, it is very important for the physicians to know the stages of aging and changes that take place during these stages and possible implications associated with them in order to more effectively treat their patients. In addition, fMRI provides information (size, density, location, etc) from brain images for the active regions during thinking and/or performing certain tasks. Thus, in this paper a monitoring brain-aging model based on a synergy of methodologies, like image segmentation, registration, local global (L-G) graphs and stochastic Petri net (SPN) graphs is presented. In particular, the synergistic brain-aging model uses fMRI images to detect, extract and associate the way that the brain regions interact regarding thinking and/or executing certain tasks. These brain-region images are extracted and geometrically are represented and associated with the L-G graphs. Then the use of SPN graphs models the regions' functionality. Thus, comparing the L-G and SPN graphs extracted from fMRI images taken in different periods from the same subject, the model has the capability to detect changes and associate them in order the medical expert to monitor the health status and provide a diagnosis or prognosis regarding a human subject. Sets of L-G and SPN graph models generated in time for each particular subject are available in an L-G/SPN graph Database. Here the synergistic model and its proof of concept are presented.

Wearable health monitoring systems (WHMS) enable ubiquitous and unobtrusive monitoring of a variety of vital signs that can be measured non-invasively. These systems have the potential to revolutionize healthcare delivery by achieving early detection of critical health changes and thus possibly even disease or hazardous event prevention. Amongst the patient populations that can greatly benefit from WHMS are Congestive Heart Failure (CHF) patients. For CHF management the detection of heart arrhythmias is of crucial importance. However, since WHMS have limited computing and storage resources, diagnostic algorithms need to be computationally inexpensive. Towards this goal, we investigate in this paper the efficiency of the Matching algorithm in deriving compact time-frequency representations of ECG data, which can then be utilized from an Artificial Neural Network (ANN) to achieve beat classification. In order to select the most appropriate decomposition structure, we examine the effect of the type of dictionary utilized (stationary wavelets, cosine packets, wavelet packets) in deriving optimal features for classification. Our results show that by applying a greedy algorithm to determine the dictionary atoms that show the greatest correlation with the ECG morphologies, an accurate, efficient and real-time beat classification scheme can be derived. Such an algorithm can then be inexpensively run on a resource-constrained portable device such as a cell phone or even directly on a smaller microcontroller-based board. The performance of our approach is evaluated using the MIT-BIH Arrhythmia database. Provided results illustrate the accuracy of the proposed method (94.9%), which together with its simplicity (a single linear transform is required for feature extraction) justify its use for real-time classification of abnormal heartbeats on a portable heart monitoring system.

Wireless capsule endoscopy is a revolutionary technology that allows physicians to examine the digestive tract of a human body in the minimum invasive way. Physicians can detect diseases such as blood-based abnormalities, polyps, ulcers, and Crohn's disease. Although this technology is really a marvel of our modern times, currently it suffers from two serious drawbacks: 1) frame rate is low (3 frames/s) and 2) no 3-D representation of the objects is captured from the camera of the capsule. In this paper we offer solutions (methodologies) that deal with each of the above issues improving the current technology without forcing hardware upgrades. These methodologies work synergistically to create smooth and visually friendly interpolated images from consecutive frames, while preserving the structure of the observed objects. They also extract and represent the texture of the surface of the digestive tract in 3-D. Thus the purpose of our methodology is not to reduce the time that the gastroenterologists need to spend to examine the video. On the contrary, the purpose is to enhance the video and therefore improve the viewing of the digestive tract leading to a more qualitative and efficient examination. The proposed work introduces 3-D capsule endoscopy textured results that have been welcomed by Digestive Specialists, Inc., Dayton, OH. Finally, illustrative results are given at the end of the paper.

Wireless Capsule Endoscopy (WCE) is a popular diagnostic technology that enables gastroenterologists to view the human digestive tract and more particularly, the small bowel, searching for various abnormalities like blood-based abnormalities, ulcers and polyps. This technology captures videos that consist of approximately 50,000 frames making its examination a very tedious task. For power consumption reasons the rate at which the frames are taken is extremely low (3 frames / second). This has a negative effect on the smoothness of the captured video and the consistency of the observed objects. This paper proposes a sophisticated video interpolation methodology for creating smooth and visually pleasing intermediate images from consecutive frames, while preserving the structure of the observed objects. It utilizes techniques and concepts from various fields such as optical flow and elasticity. Illustrative results of the methodology are also given in this paper.

Wireless Capsule Endoscopy (WCE) is an endoscopy technology that allows medical personnel to view the digestive tract non-invasively. Physicians can detect diseases such as blood-based abnormalities, polyps, ulcers and Crohn's disease. In previous papers we have proposed methodologies that deal with such abnormalities. In the current paper we are proposing a novel approach to visualize the digestive tract surface in the three-dimensional (3D) space using images (2D) of WCE videos. Current capsule technology has a low video frame rate (3 frames/second) and therefore a 3D reconstruction of the digestive tract lacks of detailed coherence. Since such a 3D reconstruction is unfeasible we focused our research on extracting and representing the texture of the surface of the digestive tract in 3D and not the digestive tract itself. The use of this technique has been well appreciated by our collaborating physicians for giving them the ability to view the WCE videos in a new different way. To our knowledge no similar research has been performed on WCE videos. Illustrative results from the methodology are also given in this paper.

The requirement to transmit video data over unreliable wireless networks is anticipated in the foreseeable future. Significant compression ratio and error resilience are both needed for applications including tele-operated robotics, vehicle-mounted cameras, sensor network, etc. Block-matching based inter-frame coding techniques, such as MPEG-x and H.26x, do not perform well in these scenarios due to error propagation between frames. Intra-only coding technologies, such as Motion-JPEG, exhibit better recovery from network data loss at the price of higher data rates. In order to address these issues, an intra-only coding scheme of H.264/AVC (iAVC) is proposed. In this approach, each frame is coded independently as an I-frame. In order to speed up the coding procedure, we propose a neural network based intra-only prediction mode decision approach, which has the potential to significantly reduce coding complexity. Frame copy is applied to compensate for packet loss. The proposed approach is a good balance between compression performance, memory usage, and error resilience. It achieves compression performance comparable to Motion-JPEG2000, with lower complexity. Low computational complexity and memory usage are very crucial to mobile stations and devices in wireless networks.

Wearable health-monitoring systems (WHMSs) represent the new generation of healthcare by providing real-time unobtrusive monitoring of patients' physiological parameters through the deployment of several on-body and even intrabody biosensors. Although several technological issues regarding WHMS still need to be resolved in order to become more applicable in real-life scenarios, it is expected that continuous ambulatory monitoring of vital signs will enable proactive personal health management and better treatment of patients suffering from chronic diseases, of the elderly population, and of emergency situations. In this paper, we present a physiological data fusion model for multisensor WHMS called Prognosis. The proposed methodology is based on a fuzzy regular language for the generation of the prognoses of the health conditions of the patient, whereby the current state of the corresponding fuzzy finite-state machine signifies the current estimated health state and context of the patient. The operation of the proposed scheme is explained via detailed examples in hypothetical scenarios. Finally, a stochastic Petri net model of the human-device interaction is presented, which illustrates how additional health status feedback can be obtained from the WHMS' user.

The last decades a variety of portable or wearable navigation systems have been developed to assist visually impaired people during navigation in known or unknown, indoor or outdoor environments. There are three main categories of these systems: electronic travel aids (ETAs), electronic orientation aids (EOAs), and position locator devices (PLDs). This paper presents a comparative survey among portable/wearable obstacle detection/avoidance systems (a subcategory of ETAs) in an effort to inform the research community and users about the capabilities of these systems and about the progress in assistive technology for visually impaired people. The survey is based on various features and performance parameters of the systems that classify them in categories, giving qualitative-quantitative measures. Finally, it offers a ranking, which will serve only as a reference point and not as a critique on these systems.

The design and development of wearable biosensor systems for health monitoring has garnered lots of attention in the scientific community and the industry during the last years. Mainly motivated by increasing healthcare costs and propelled by recent technological advances in miniature biosensing devices, smart textiles, microelectronics, and wireless communications, the continuous advance of wearable sensor-based systems will potentially transform the future of healthcare by enabling proactive personal health management and ubiquitous monitoring of a patient's health condition. These systems can comprise various types of small physiological sensors, transmission modules and processing capabilities, and can thus facilitate low-cost wearable unobtrusive solutions for continuous all-day and any-place health, mental and activity status monitoring. This paper attempts to comprehensively review the current research and development on wearable biosensor systems for health monitoring. A variety of system implementations are compared in an approach to identify the technological shortcomings of the current state-of-the-art in wearable biosensor solutions. An emphasis is given to multiparameter physiological sensing system designs, providing reliable vital signs measurements and incorporating real-time decision support for early detection of symptoms or context awareness. In order to evaluate the maturity level of the top current achievements in wearable health-monitoring systems, a set of significant features, that best describe the functionality and the characteristics of the systems, has been selected to derive a thorough study. The aim of this survey is not to criticize, but to serve as a reference for researchers and developers in this scientific area and to provide direction for future research improvements.

In this work, we present a local-global (LG) graph methodology for iris based biometric authentication. Local-global (LG) graph method adds local part information into a global graph. Local graphs of the pre-processed iris images are first calculated by feature extraction and combined to form a global graph that is stored in a database for the purpose of authentication. The global graph of the presented test image is compared with the global graph of the stored reference image and based on a distance metric, the authenticity of the subject is established.

While reading devices for the visually impaired have been available for many years, they are often expensive and difficult to use. In an effort to prototype a device comprised of inexpensive components, we present a highly interactive method for processing document images. The voice user interface (VUI), document image segmentation and framing, and targeted image registration methods of a mobile reading device are presented. Initial experimental results on a set of newspaper pages are also presented.

In this paper we present an enhanced version of our NRF spectra classifier based on local-global graphs (LG-graphs) and provide a comparison with other possible alternative detection schemes. Experimental results verify our claims that the proposed nuclear resonance fluorescence (NRF) signature detection methodology is favorable over other widely employed conventional feature extraction and signal representation methods. The LG-graph methodology is based on the representation of the signal's peaks as triangle-like shapes and then on extracting significant geometrical features from these triangles to derive a concise representation of the peaks that correspond to specific NRF signatures of interest. These features are used to enable the matching of the materials of interest in a new unknown NRF spectrum.

The deployment of wearable health monitoring systems (WHMS) is expected to address several important healthcare-related issues such as increasing healthcare costs, the rising number of the elderly population and treatment of chronic conditions. However, most of the currently developed WHMS simply serve as ambulatory physiological data loggers and transmitters in order to make the recorded bio-signals remotely available for inspection from a supervising physician. In this paper we describe our efforts towards setting-up a WHMS prototype that is capable of providing individualized embedded decision/diagnosis support for round-the-clock remote health monitoring of people at risk. To realize this goal an ANN-based approach is adopted, whereby a supervised learning period is required in order to embed patient-specific medical knowledge into the system, which will then enable it to make more accurate and ¿safer¿ estimations about the user's health condition.

As the number of older Americans increases and many decide to stay in their homes, the need for assistive technologies grows. One such technology is an intelligent system of surveillance cameras. While these systems can provide many services, we attempted to limit ourselves to the evaluation of systems that use cameras to track the locations of objects that may be lost or misplaced within a home. We provide a comparison of the systems that are currently in operation or are being developed and suggest areas where enhancements may produce systems that are more effective. Because few systems exist for the sole purpose of tracking objects in a home, we also evaluate systems that are designed to use multiple cameras to perform general surveillance. We go on to show where some principles of these system can be adapted to improve the performance in the task of tracking objects within a home.

This work describes two new pause detection algorithms and compare their performance with four standard Voice Activity Detection (VAD) methods represented by the adaptive Long Term Spectral Divergence (LTSD) algorithm, the Likelihood Ratio Test (LRT) algorithm, the Neural Network thresholding and G.729. The proposed algorithms exploit the concept of adaptation in order to handle adverse conditions and spontaneous speech properties. The test data are recordings of spontaneous speech made in noisy environments. The experimental results show that the performance of proposed algorithms on noisy and even artificially cleaned speech are superior than that achieved by standard methods reported in literature .

In this paper we present research work towards a 2D tactile vocabulary for training visually impaired for their independent mobility. The vocabulary is associated to a 2D tactile array (vibration array), which is a part of a wearable navigation prototype (Electronic Travel Aid) called Tyflos. The vibration array is currently consisting of 16 vibrating elements arranged in a 4×4 manner. Each motor can be independently driven with square pulses of varying frequencies. The vibration array can represent dynamically the 3D space of the user's field of view: the 2D arrangement represent the x-y coordinates while the frequencies represent the z coordinate (which is the distance of an obstacle from the user). Different navigation and human factor criteria have been used to create the 2D tactile vocabulary. Finally, using the continuous feedback from the users, the goal is to balance between a minimal vocabulary for easy learning and a rich vocabulary that will still be able to represent efficiently the 3D navigation space.

This paper deals with a novel methodology for automatically detecting blood based abnormalities (bleeding, angioectasia and erythema) in wireless capsule endoscopy videos. The methodology is based on a synergistic integration of methods, such as Color K-L transformation, fuzzy region segmentation, Local-Global graphs. The methodology presents several unique features that separate it from the classical bleeding based methods. These features are mainly based on (i) the classification of the blood based pathological cases in three major categories approved by gastroenterologists; (ii) it is based on a synergy of methods tested in the research field for their robustness. Illustrative examples from the methodology are also presented in this paper.

Wearable health-monitoring systems (WHMS) promise to revolutionize health care by providing real-time unobtrusive monitoring of patientspsila physiological parameters through the deployment of several on-body and even intra-body biosensors. Although several technological issues regarding WHMS still need to be resolved, in order for them to become more applicable in real-life scenarios, it is expected that continuous ambulatory monitoring of vital signs will enable pro-active personal health management and better treatment of patients suffering from chronic diseases, of the elderly population and of emergency situations. In this paper a novel formal language based model for multi-sensor data fusion and early-detection of various conditions is presented. Patterns or even signal states indicating pathological symptoms that are presented in the signals, which can be collected from on-body distributed biosensors, are modeled as symbols of the Prognosis context-free formal language, whose grammar and production rules define the prognosis-words. The proposed approach is based on a described generic WHMS model and on a simple but at the same time efficient method for characterizing body-signalpsilas patterns and/or states. Finally, we provide several illustrative examples for better comprehension of the proposed model.

The efficient processing and association of different multi-modal information is a very important research field with a great variety of applications, such as human computer interaction, knowledge discovery, document understanding, etc. A good approach to this important issue is the development of a common platform for converting different modalities (such as images, text, etc) into the same medium and associating them for efficient processing and understanding. Thus, this paper here presents the development of a novel methodology based on Local-Global (LG) graphs capable for automatically converting image context into natural language text sentences and then into speech for serving as an interactive model for locating missing objects in home environments. Simple illustrative examples are provided for proving the concept proposed here.

This paper presents the multimodal interaction scheme (visual and audio) used by the Tyflos prototype. Tyflos is a wearable prototype that provides reading and navigating assistance for visually impaired users. In particular, the Tyflos prototype integrates a wireless portable computer, cameras, range and GPS sensors, microphones, natural language processor, text-to-speech device, an ear speaker, a speech synthesizer, a 2D vibration vest and a digital audio recorder. Data collected by the Tyflos sensors is processed by appropriate modules, each of which is specialized in one or more tasks. In this paper we also present a Stochastic Petri-net model of the multimodal interaction scheme for both of the Tyflos capabilities, reading and navigation. Simple illustrative examples from reading and navigation cases are also presented to demonstrate the multimodal interaction.

The objective of this paper is to develop an NRF signature classifier scheme in order to efficiently and accurately detect and recognize NRF signatures of nuclear material inside a cargo container. In response to this problem we have developed a local-global (LG) graph as the NRF signal representation scheme for efficient matching and on which to build the NRF signature library. Here we also assume that the NRF signals will be "free" of noise after passing a preprocessing step at a previous stage. Thus, the proposed scheme is based on the analysis of an incoming NRF signal and determining its min and max; Then we create the local triangles and express them in a form of local graphs; at the last step we combine local graphs for creating the Local-Global graphs that hold the desirable pattern (signature) of certain material.

We consider the problem of recognizing 3-D objects from 2-D images using geometric models and assuming different viewing angles and positions. Our goal is to recognize and localize instances of specific objects (i.e., model-based) in a scene. This is in contrast to category-based object recognition methods where the goal is to search for instances of objects that belong to a certain visual category (e.g., faces or cars). The key contribution of our work is improving 3-D object recognition by integrating Algebraic Functions of Views (AFoVs), a powerful framework for predicting the geometric appearance of an object due to viewpoint changes, with indexing and learning. During training, we compute the space of views that groups of object features can produce under the assumption of 3-D linear transformations, by combining a small number of reference views that contain the object features using AFoVs. Unrealistic views (e.g., due to the assumption of 3-D linear transformations) are eliminated by imposing a pair of rigidity constraints based on knowledge of the transformation between the reference views of the object. To represent the space of views that an object can produce compactly while allowing efficient hypothesis generation during recognition, we propose combining indexing with learning in two stages. In the first stage, we sample the space of views of an object sparsely and represent information about the samples using indexing. In the second stage, we build probabilistic models of shape appearance by sampling the space of views of the object densely and learning the manifold formed by the samples. Learning employs the Expectation-Maximization (EM) algorithm and takes place in a "universal," lower-dimensional, space computed through Random Projection (RP). During recognition, we extract groups of point features from the scene and we use indexing to retrieve the most feasible model groups that might have produced them (i.e., hypothesis generation). The likelihood of each hypothesis is then computed using the probabilistic models of shape appearance. Only hypotheses ranked high enough are considered for further verification with the most likely hypotheses verified first. The proposed approach has been evaluated using both artificial and real data, illustrating promising performance. We also present preliminary results illustrating extensions of the AFoVs framework to predict the intensity appearance of an object. In this context, we have built a hybrid recognition framework that exploits geometric knowledge to hypothesize the location of an object in the scene and both geometrical and intesnity information to verify the hypotheses.

This paper deals with the design, simulation and implementation of a 2D vibration array used as a major component of an assistive wearable navigation device for visual impaired. The 2D vibration array consists of 16 (4x4) miniature vibrators connected to a portable computer, which is the main computing component of the entire wearable navigation system, called Tyflos. Tyflos consists of two miniature cameras (attached to a pair of dark glasses), a microphone, an ear speaker, the 2D vibration array, and a portable computer. The cameras capture images from the surrounding environment and after appropriate processing 3D representations are created. These 3D space representations are projected on the 2D array, which vibrates in various levels corresponding to the distances of the surrounding obstacles. The 2D array is attached to the user's chest in order to provide the appropriate sensation (via vibrations) of the distances from the surroundings.

SCAN is a class of formal languages for compression, encryption and information hiding. We have previously studied and reported separate hardware implementations of SCAN compression and encryption. This paper presents initial results on the design of a complete single-chip system for the SCAN compression, encryption and information hiding algorithm using the stretch technology with reconfigurable and fixed resources. The result is a simple, low cost, embeddable core combining all three operations seamlessly and the design was fully mapped to the stretch technology.

In a multimedia information system, different types of digital media (such as video, audio, etc.) are stored, transmitted, and presented. During presentation time, the synchronization between audio and video data has to be preserved in order to offer the best perceptual quality. However, the timestamp-based synchronization methodology in MPEG/System layer suffers data packet loss during transmission, and the resulted absence of synchronization will be unacceptable for users. In order to address this issue, an information hiding based synchronization methodology has been proposed. In this methodology, audio data is embedded within the corresponding video frames by means of high bitrate information hiding techniques. At the receiver, the embedded audio data is extracted and played with the host video frames to achieve the synchronization. With this approach, significant advantages have been obtained: (1) the communication channel for audio data transmission is avoided; (2) the synchronization between audio and video data is robust to packet loss.

One important and challenging problem in human interaction is the communication between blind and deaf individuals. The challenge here involves several cases: (i) first case is a deaf person usually does not speak in order a blind person to hear him/her; (ii) second case is when a blind person speaks a deaf person cannot hear; (iii) third case is when a deaf person makes sign language signs a blind person cannot see them. Thus, this paper presents a study on multi-modal interfaces, issues and problems for establishing communication and interaction between blind and deaf persons. A system-prototype Tyflos-Koufos is proposed in an effort for offering solutions to these challenges.

This paper presents a methodology that uses formal languages representation for image processing for automatically detecting texture patterns and determining texture paths in images. The texture paths are detected and extracted by using image processing techniques, such as image segmentation to isolate regions of interest, and then the extraction of repeating textures. The detection of the texture blocks is obtained by recursively using 27 times 27, 9 times 9, and 3 times 3 windows. Predefined repeating texture patterns are also searched for in each of the sets. For each set of texture blocks with similar or same characteristics, curve fitting techniques are used for association of patterns with the texture. The selected curve is split into straight line segments, and the pattern is finally represented using the defined context-free formal language methodology. The methodology has capabilities to learn texture paths associated with certain "objects", which generate them, and use this knowledge for a variety of applications. Results are shown for various color images.

The goal of image chromatic adaptation is to remove the effect of illumination and to obtain color data that reflects precisely the physical contents of the scene. We present in this paper an approach to image chromatic adaptation using Neural Networks (NN) with application for detecting--adapting human skin color. The NN is trained on randomly chosen color images containing human subject under various illuminating conditions, thereby enabling the model to dynamically adapt to the changing illumination conditions. The proposed network predicts directly the illuminant estimate in the image so as to adapt to human skin color. The comparison of our method with Gray World, White Patch and NN on White Patch methods for skin color stabilization is presented. The skin regions in the NN stabilized images are successfully detected using a computationally inexpensive thresholding operation. We also present results on detecting skin regions on a data set of test images. The results are promising and suggest a new approach for adapting human skin color using neural networks.

"There are million individuals with disabilities with traumatic emotional experiences due to their health issues leading frustration and depression, where a major factor for it is their emotional behavior". Emotion is a topic that has received much attention during the last few years, both in the context of speech synthesis, image understanding as well in automatic speech recognition, interactive dialogues systems and wearable computing. There are few promising studies on the emotional behavior of people with disabilities. These studies are partial due to the lack Information technology and engineering (ITE) techniques that make available a deeper and large scale non-invasive analysis and evaluation of the disabled people emotional behavior in order to provide tools and support for helping them to overcome social and health barriers. A quantitative and qualitative study of emotional invariant meta-features to support the development of emotionally rich man-machine interfaces (interactive dialogue systems and intelligent avatars) for people with disabilities is the subject of this paper

Phoneme recognition strongly depends on the intrinsic duration of speech segments, phoneme spectral change's durations, and the relative timing of two overlapping events. Excerpts of fluent speech are not very well perceived below a given duration threshold, and "phonemic clauses" are signalled either by a speech pause, or the lengthening of the final word or syllable in the clause. If a sentence should be perceived as temporally fluent, changes made in the duration of one segment should be compensated by durational changes of adjacent segments. These data lead to the conclusion that temporal aspects have a primary role in vocal communication and their perception is basic for a correct exchange of information and a correct identification of the phonetic and phonologic characteristics of the vocal message. In the light of these considerations, the role that the perception of temporal phenomena plays in learning or reading language is investigated considering two theories proposed to explain reading and specific language impairments. The first theory assumes that poor readers are affected by general auditory deficits in temporal processing, whereas the second assumes that the above impairments arose from specific deficits in learning speech. Experimental data in favour of the first or the second theory are discussed, and at the light of the reported results, new experimental paradigms are suggested

Representing, associating, interpreting and learning behavioral patterns and rules from observation is always a challenging research problem for cognitive systems and human computer interaction. This problem becomes more difficult in complex environments with many unexpected behavioral patterns from objects, or cells, or individuals or small group of people acting under certain plans at the same time in different locations. This paper targets the development of a novel synergistic methodology for automatically monitoring, representing, associating and interpreting behavioral patterns and activities, and learning rules from complex environments using videos

In this article, we present a local global graph (LGG) method for recognizing facial expressions from static images irrespective of different illumination conditions, shadows and cluttered backgrounds. First, a neural color constancy based skin detection procedure to detect skin in complex real world images is presented. Second, the LGG method for detecting faces and facial expressions with a maximum confidence from skin segmented images is presented. The LGG approach presented here emulates the human visual perception for face and expression detection. In general, humans first extract the most important facial features such as eyes, nose, mouth, etc. and then inter-relate them for face and facial expression representations. The LG Graph embeds both the local information (the shape of facial feature is stored within the local graph at each node) and the global information (the topology of the face). Facial expression recognition from the detected face images is obtained by comparing the LG Expression Graphs with the existing the LG expression models present in the LGG database. Experimental results on the AR database and real-world images suggest the robustness of the proposed approach for facial expression recognition

In this paper, we present an image-text dewarping methodology based on robust estimation of text-lines. When a text-page is captured by a camera, it suffers both from the perspective distortions and the page curl. The non-linear distortion due to page-curl is inherently present, given the surface nature of the pages and the text-book. The state-of-the-art OCR systems have a very low performance on recognizing such distorted text. To remove both these distortions and to produce a flattened view of the text, we use the cues present in the image-text, i.e., the text-lines on the surface of the page are straight. The methodology requires only a single camera captured image and does not require any calibration or any other expensive hardware setups as in other methods. Experimental results on a set of documents show that the methodology produces visually pleasing output and also improves OCR accuracy

This paper presents a methodology for detecting abnormal patterns in wireless capsule endoscopy (WCE) images. In particular, an average of 50,000 images are obtained during an WCE exam. Usually, these images are reviewed in a form of a video at speeds between 5 to 40 image-frames/sec. The time spent by a physician reading the results of WCE images varies between 45 to 180 minutes. This presents a major problem which is the reading process that consumes a significant amount of time and the results take several days before they become available since the physician has to find the time to study each video uninterrupted for up to 3 hours. The methodology presented here is based on the automatic detection of abnormal WCE patterns in an effort for reducing the reading time of the WCE images and the cost of the procedure as well. The methodology consists of a synergistic integration of image processing, analysis and recognition techniques for achieving the automatic detection of the WCE abnormal patterns.

This paper presents a 2D vibration array for detecting dynamic changes in 3D space during navigation and provides these changes in real-time to visual impaired users (in a form of vibration) in order to develop a 3D sensing of the space and assist their navigation in their working and living environment. This vibration array is a part of the Tyflos prototype device (consisting of two tiny cameras, a microphone, an ear speaker mounted into a pair of dark glasses and connected into a portable PC) for blind individuals. The overall idea is of detecting changes in a 3D space is based on fusing range data and image data captured by the cameras and creating the 3D representation of the surrounding space. This 3D representation of the space and its changes are mapped onto a 2D vibration array placed on the chest of the blind user. The degree of vibration offers a sensing of the 3D space and its changes to the user.

This paper deals with development of a methodology for detecting bleeding in WCE images. The presented methodology is based on a neural net model available in the AIIS Inc. and the results of this methods were tested at the AT research lab. The performance of our method offers promising results in comparison with the ones made by the given imaging RBIS (red blood identification system), which vary from 21% to 41%. Our method is under improvements and the expected results will reach near 80% or more.

The SCAN methodology is a class of formal languages which can be applied to compression, encryption, and data hiding or in combinations thereof. The SCAN algorithm for video compression belongs to the category of interframe techniques. Each frame is compared to its preceding one and through frame differencing redundant information is removed. As in all algorithms of this category key frames are placed as beginning points for finding the differences with subsequent frames. Frames between key frames are named difference frames and they only contain the information that has changed from the previous frame. The differences between frames are determined using a user-defined threshold value, and therefore in general this type of compression can be lossy. The consequent degradation of picture is controlled by the user via the choice of threshold value. When this threshold is 0 the compression becomes lossless. The final goal is to develop a seamless architecture for compression, encryption and information hiding.

A multiresolution color image segmentation approach is presented that incorporates the main principles of region-based segmentation and cluster-analysis approaches. The contribution of This work may be divided into two parts. In the first part, a multiscale dissimilarity measure is proposed that makes use of a feature transformation operation to measure the interregion relations with respect to their proximity to the main clusters of the image. As a part of this process, an original approach is also presented to generate a multiscale representation of the image information using nonparametric clustering. In the second part, a graph theoretic algorithm is proposed to synthesize regions and produce the final segmentation results. The latter algorithm emerged from a brief analysis of fuzzy similarity relations in the context of clustering algorithms. This analysis indicates that the segmentation methods in general may be formulated sufficiently and concisely by means of similarity relations theory. The proposed scheme produces satisfying results and its efficiency is indicated by comparing it with: 1) the single scale version of dissimilarity measure and 2) several earlier graph theoretic merging approaches proposed in the literature. Finally, the multiscale processing and region-synthesis properties validate our method for applications, such as object recognition, image retrieval, and emulation of human visual perception.

The goal of image chromatic adaptation is to remove the effect of illumination and to obtain color data that reflects precisely the physical contents of the scene. We present An approach to image chromatic adaptation using neural networks (NN) with application for detecting - adapting human skin color. The network is trained on randomly chosen color images containing human subject under various illuminating conditions, thereby enabling the model to dynamically adapt to the changing illumination conditions. The proposed network predicts directly the illuminant estimate in the image so as to adapt to the human skin color. The comparison of our method with gray world, white patch and neural network on white patch algorithms is presented. We also present our results on detecting skin regions in NN color corrected test images. The results are promising and suggest a new approach for adapting human skin color using NN's.

The analysis of motion in image sequences represents a popular and constantly growing research field. In This work a motion analysis scheme is proposed to function as an assistive vision technology tool. The first stage of this methodology includes a video stabilization scheme as a preprocessing stage. In the next stage a nonlinear spatio-temporal diffusion approach is applied and a kernel based density estimation method follows to assess the motion activity. This overall process is completed by watershed-based segmentation to detect the moving areas. The presented experimental results indicate the efficiency of this scheme in detecting and segmenting the moving areas.

The topic of this work is related to image registration of aerial images. This topic represents a very complicated problem especially when several variations and distortions occur. In addition to that, the requirement for an automated process further complicates this task. In this paper a stochastic optimization scheme is proposed using genetic algorithms to address misalignments caused by viewpoint, temporal and terrain variations of aerial images. A multiscale optical flow approach is applied next to achieve subpixel registration. Some experimental results are also presented to indicate the applicability of the proposed scheme.

This paper briefly presents a block cipher encryption architecture and a reconfigurable logic-based hardware design for the SCAN encryption algorithm. Detailed performance results are presented for still images as well as video, and the reconfigurable architecture is compared to software-only implementations of the same algorithm as well as a preliminary ASIC design.

This paper presents a brain based biometrics methodology using local global (L-G) and stochastic Petri net (SPN) graphs. In particular, today's most used lying detection systems (polygraph) are based on blood pressure under certain psychological conditions and have been proven inadequate and inaccurate for determining if a person is saying the truth or not. The main reason behind this inaccuracy is the psychological conditions under the test is contacting. Innocent persons with psychological fear to police or FBI agents may inappropriately react during the test producing false alarms, while at the same time guilty persons with acting capabilities may be lying and pass the test. Here in this paper we present a methodology based on fMRI and the way that the brain cells or brain regions interact regarding thinking and/or executing certain tasks. More specifically, fMRI provides color images for the active regions of the brain during thinking and/or performing certain tasks. These region-images are extracted and geometrically (topologically) are associated with the L-G graphs. Then using SPN graphs we represent their functionality. Thus, comparing the L-G and SPN graphs extracted from the candidate fMRI images to a set of "correct" generic L-G and SPN graph models available in an L-G/SPN graph database offers a more accurate measurement of the truth of false answers during the test. The main reason about the accuracy is that a liar will be detected because he/she will try to "think" how to avoid the truth.

This paper presents a methodology for object recognition using the local-global (L-G) graph approach. In particular, the L-G graph approach based on an image segmentation approach and synthesis of the regions that compose an object. The synthesis of regions is based on the L-G modeling that compares a set of L-G based object models sitting in an L-G Database. The methodology accurate for objects existed in the DB and it has the ability of "learning". Illustrative examples are also provided.

This paper deals with a 3D methodology for brain tumor image-guided surgery. The methodology is based on development of a visualization process that mimics the human surgeon behavior and decision-making. In particular, it originally constructs a 3D representation of a tumor by using the segmented version of the 2D MRI images. Then it develops an optimal path for the tumor extraction based on minimizing the surgical effort and penetration area. A cost function, incorporated in this process, minimizes the damage surrounding healthy tissues taking into consideration the constraints of a new snake-like surgical tool proposed here. The tumor extraction method presented in this paper is compared with the ordinary method used on brain surgery, which is based on a straight-line based surgical tool. Illustrative examples based on real simulations present the advantages of the 3D methodology proposed here.

This correspondence presents the basic design and the simulation of a low level multilayer vision processor that emulates to some degree the functional behavior of a human retina. This retina-like multilayer processor is the lower part of an autonomous self-organized vision system, called Kydon, that could be used on visually impaired people with a damaged visual cerebral cortex. The Kydon vision system, however, is not presented in this paper. The retina-like processor consists of four major layers, where each of them is an array processor based on hexagonal, autonomous processing elements that perform a certain set of low level vision tasks, such as smoothing and light adaptation, edge detection, segmentation, line recognition and region-graph generation. At each layer, the array processor is a 2D array of k/spl times/m hexagonal identical autonomous cells that simultaneously execute certain low level vision tasks. Thus, the hardware design and the simulation at the transistor level of the processing elements (PEs) of the retina-like processor and its simulated functionality with illustrative examples are provided in this paper.

This paper presents two methods for comparison of images and evaluation of visibility of artifacts due to hidden information, changes or noise. The first method is based on pixel flow functions (PFF) able to detect changes in images by projecting the pixel values vertically, horizontally and diagonally. These projections create "functions" related with the average values of pixels summarized horizontally, vertically and diagonally. These functions represent image signatures. The comparison of image signatures defines differences in images. The second method is based on a heuristic graph model, known as local-global graph (LGG), for evaluating visibility of modifications in digital images. The LGG is based on segmentation and comparing the segments while thresholding the differences in their attributes. The methods have been implemented in C++ and their performance is presented.

The paper focuses on the development of an efficient and accurate tool for segmenting color images. The segmentation is a problem that has been widely studied since machine vision first evolved as a research area. The neural network segmentation tools and technology developed and presented in this paper show great potential in application where the accuracy is the major factor. Similar requirements exist in the area of medical imaging where segmentation must provide the highest possible precision. The feasibility of the work presented shows a promising future by using a cluster-based approach to training very large feedforward neural networks.

This paper presents a new methodology for efficiently representing the content of images and comparing images by detecting and recording their visual differences. In particular, the methodology presented here is based on a stochastic Petri-net (SPN) graph approach able to extract and record local and global features from both images, compare them, and define the percentage of similarity. One of the features of the human visual perception is the detection of similarities between two images. The visual similarity is based on color, size, shape, and local and global topological changes of the image regions. Several methods dealing with image or object similarities have been proposed. The new feature of the method here is the partial emulation of the human observer's visual perception by recording differences extracted from different images. Results of the method described here are presented for a variety of images by using local and global noisy conditions.

This paper presents a method of knowledge representation for very large scale integration (VLSI) chip design which provides the necessary information for abstraction from the physical design to gate-level logic through a high-level behavioral model. The representation scheme used by the ANTISTROFEAS system utilizes a hierarchical attributed graph structure which consists of incrementally abstracted design information for the VLSI system. This method of knowledge representation is well-suited to reverse-engineering of VLSI chips from the layer mask layout data, but is also applicable to applications at many levels of the design process including design rule checking, logic synthesis, design verification, and partitioning-compaction problems. The representation scheme is applicable to any VLSI technology, and is designed to take advantage of artificial intelligence. expert system techniques, by disassociating the representation and manipulation of the VLSI design data from the rules which govern its correctness and transformation for other usage.

This paper presents the navigation methodology employed by an intelligent assistant (agent) for people with disabilities. In particular, the intelligent assistant, called Tyflos, would help a visually impaired user to be partially independent and able to walk and work in a 3D dynamic environment. The Tyflos system carries two vision cameras and captures images from the surrounding 3D environment, either by the user's command or in a continuous mode (video), then it converts these images into verbal descriptions for each image into a verbal communication with the user. In other words the system plays the role of human assistant, who describes to the user the 3D visual environment.

This paper presents a study on the behavior of smart cars by considering them as autonomous intelligent agents. In particular, a smart car could behave as autonomous agent by extracting information from the surrounding environment (road, highway) and determining its position in it, detecting the motion and tracking the behavioral patterns of other moving objects (automobiles) in its own surrounding space, exchanging information via internet with other moving objects (if possible) and negotiating its safety during travel with the other moving objects.

Presents the basic design and the simulation of a low level multi-layer vision processor that emulates to some degree the functional behavior of a human retina. This retina-like multi-layer processor is the lower part of an autonomous self-organized vision system, called Kydon, which could be used on visually impaired people with a damaged visual cerebral cortex. The retina-like processor consists of four major layers, where each is an array processor that performs a certain set of low level vision tasks, such as smoothing and light adaptation, edge detection, segmentation, line recognition and region-graph generation. At each layer the array processor is a 2-D array of kxm hexagonal identical autonomous cells that simultaneously execute certain tasks. Thus, in this paper the hardware design and the entire simulation of the retina-like processor with illustrative examples are provided.

This paper describes a new approach for representing, recognizing and interpreting human activity from video. The approach presented (at the conceptual level) is a model based on the hierarchical synergy of three other models (the L-G graph, the SPN graph and a NN model). In particular, in our project human activity is strongly related with the ability of describing and interrelating events. Thus, the L-G graph (local-global graph) provides a powerful description of the structural image features presented in an event, the SPN (stochastic Petri net) model offers a description of the functional behavior of the changes or operations in video presented in an event, and the NN (neural network) model provides the capability of extracting and learning behavioral patterns, presented in human activities.

This paper presents a software tool-methodology called ORASIS, for near real-time and real-time user driven segmentation of color images or sequences of images (video) by using a low cost approach (a low cost camera and a PC). The segmentation approach used is based on fuzzy-like reasoning and a color reflection. The ORASIS tool-methodology provides a human-like perception modeling in combination with fast smoothing, region growing and edge detection. Results on several images are provided to demonstrate the tool's adaptive segmentation capabilities based on the user's requests.

This paper deals with the development of 2-D space maps during navigation in unknown environments, by synthesizing the shapes of consecutive free spaces. Each shape extracted from each current-free navigation space is represented by attributed graphs. Thus, the synthesis of the shapes is based on the synthesis of graphs for the generation of new graph forms that represent the new shape traveled by an autonomous robot. Illustrative examples from a space with obstacles are presented.

An efficient first grasp for a wheelchair robotic arm-hand with pressure sensing is determined and presented. The grasp is learned by combining the advantages of neural networks and fuzzy logic into a hybrid control algorithm which learns from its tip and slip control experiences. Neurofuzzy modifications are outlined, and basic steps are demonstrated in preparation for physical implementation. Choice of object approach vector based on fuzzy tip and slip data and an expert supervisor, as well as training of a diagnostic neural tip and slip controller, are the focus of this work.

The paper presents a methodology for document processing and understanding. The methodology assumes that text paragraphs have been separated from the text images and that the necessary interrelationships are for a possible reconstruction of the original page. Each text line forms a natural language text which requires analysis in order to contribute to text understanding. The analysis is based on the formal representation and processing of natural language text by using stochastic Petri nets (SPNs). Text understanding is based on the semantic analysis/synthesis of SPN forms for the appropriate interpretations. These interpretations in combinations with image descriptions and/or objects extracted from images contribute to document understanding.

This paper describes a methodology for removing (partially or totally) redundant information received from different documents in an effort to synthesize new documents. In particular, information retrieved from different databases may have various forms, such as images, natural language text, data, etc. These pieces of information may be parts of one or more documents related with a specific subject. This means that a number of text-paragraphs and images may occur (or retrieved) more than once, by creating redundancy in the storage space. Thus, in order to create a new redundant-less document the duplicated parts of information have to be removed. The methodology presented analyzes text-paragraphs and images received from different DBs by using a set of similarity criteria in order to make a decision for the removal of the duplicated ones. Illustrative examples are provided.

This paper deals with the modeling of an extended quartet multiprocessor kernel used for the evaluation of the Hermes system. Hermes is a multiprocessor hybrid system architecture used as a machine vision system. The functionality of Hermes requires an asynchronous information flow upwards and downwards, where "orders"(in a form of code and data) go down and "abstracted" or processed picture information goes up along the system's hierarchy. Moreover, the overall functional behavior of the Hermes system can be considered as an extended quartet kernel in an abstracted manner. The extended quartet kernel studied here presents either a memory-to-memory (M-M) or a bus-to-bus (B-B) connectivity. The evaluation of these two kernel's configurations (M-M, B-B) defines that the structural design of the Hermes system will be based on the M-M scheme. When failures occur on the quartet kernel, however, the M-M scheme will be converted into a B-B one by using a failure recovery procedure. The evaluation of the quartet kernel configurations is based on a probabilistic model. Failures and recovery procedures on the quartet kernel are discussed and the performance evaluation of the kernel (under failures) is also provided.

Edge detection is one of the most important image processing steps towards image understanding. It is desired that edges be continuous and that the resultant regions or segments be completely isolated from their neighbors. Initially, images must first be smoothed to remove noise. In this paper, a novel fuzzy-like smoother algorithm is presented which removes camera noise and enhances edge contrast. The edge detection algorithm, which is applied on the smoothed image, is then presented. In this algorithm normalized hue in HSI space and color contrast in RGB space are combined using an aggregate operator. Pixels considered to be at least "nearly" locally maximum (defined within) are then found for all edge directions and the results are combined.

Neurofuzzy approaches for predicting financial time series are investigated and shown to perform well in the context of various trading strategies involving stocks and options. The horizon of prediction is typically a few days and trading strategies are examined using historical data. Two methodologies are presented wherein neural predictors are used to anticipate the general behavior of financial indexes (moving up, down, or staying constant) in the context of stocks and options trading. The methodologies are tested with actual financial data and show considerable promise as a decision making and planning tool.

In this paper a fusion technique is presented for 3-D representation of space. The technique is based on a fusion process which involves laser range data and segmented image data with human expertise related with the surrounding environment. In particular, the range data may contain noise due to reflections on sloped surfaces or long distance open corridors, thus the proposed fusion approach removes the noise by using human expertise from 3-D environments. This method could be utilized either by an autonomous robot in an unknown environment, or by an inspection machine in a complex manufacturing environment, or by a visual navigation device used by blind people. Additional applications for this technique are the ability to correct measurement deficiencies in a laser-scan and to provide a true color and 3-D perceived shape representation for a given object in a modeling environment.

Presents a methodology for visually tracking and extracting targets from a sequence of images (video). The methodology presented in this paper consists of a combination of algorithms, such as heuristic segmentation, edge detection, thinning, region growing, fractals, feature extraction, graphs with attributes, etc., appropriately selected according to the existing situation, such as moving target/still camera, still target/moving camera and moving target/moving camera. The new contribution of this paper is the combination of algorithms in a human-like feedback geometric approach for processing low-resolution information from consecutive images. Simulated results of the methodology are presented.

This paper presents a generic traffic priority language, called KYKLOFORTA, used by autonomous robots for collision-free navigation in a dynamic unknown or known navigation space. In a previous work by X. Grossmman (1988), a set of traffic control rules was developed for the navigation of the robots on the lines of a two-dimensional (2-D) grid and a control center coordinated and synchronized their movements. In this work, the robots are considered autonomous: they are moving anywhere and in any direction inside the free space, and there is no need of a central control to coordinate and synchronize them. The requirements for each robot are i) visual perception, ii) range sensors, and iii) the ability of each robot to detect other moving objects in the same free navigation space, define the other objects perceived size, their velocity and their directions. Based on these assumptions, a traffic priority language is needed for each robot, making it able to decide during the navigation and avoid possible collision with other moving objects. The traffic priority language proposed here is based on a set of primitive traffic priority alphabet and rules which compose pattern of corridors for the application of the traffic priority rules.

This paper describes the VLSI design and simulation of the lower layer processors of the KYDON vision system. KYDON is a completely autonomous, hierarchical, multilayered image understanding system. The VLSI design of the individual components as well as the timing simulation results of the processor array have been presented. The system runs at 50 MHz and promises a high processing rate of 300 image frames/sec.

We study the success of a neural network computer model to predict the price, of a stock, given the fluctuations in the rest of the market that day. Based on the neural net's prediction, the program then measures its success by simulating buying or selling that stock, based on whether the market's price is determined over valued or under valued. The neural net itself is a modification of a fuzzy based neural network (Kung, 1993). The program consistently averages over 20% A.P.R. and has been time tested over 6 years with several stocks.

In this paper, a new efficient methodology for building block layout is presented by using synthesis placement and compaction. The synthesis placement part of the methodology is based on a formal language called GEOMETRIA. The compaction part is based on geometric reshapings (gs) of blocks and the merging of the communication channels. Both reshaping and merging follow the VLSI regulations for legal layout placement and improve the overall functional performance of the integrated system, by reducing the average length of the connection lines and the size of the occupied chip area by retaining the functionality and the neighboring relations of the blocks. The main goal of the blocks' geometric reshaping is minimization of the wasted area (or dead space among the blocks) called "open holes". The channels merging process of compaction is based on the legal overlapping of the blocks' communication channels by reducing the layout placement at the local and global routing.

This paper presents a parallel thinning algorithm which processes binary images and produces the image skeletons which are either regions size independent (noise non-sensitive in small variation of shape (PSTA2)), or noise sensitive to the regions shape (PSTA1). A comparison of the proposed algorithm (parallel symmetric thinning algorithm (PSTA)) to a number of other parallel thinning algorithms is also provided indicating its advantages and disadvantages.

In this paper, a learning model for an autonomous vision multi-layer architecture, called KYDON, is presented modeled and analyzed. This learning model uses a birth and death approach to derive the relationships among the parameters used in the learning characteristic function. In addition the two critical (deletion and saturation) points on the learning curve are evaluated. These points represent two extreme states on the learning process. The KYDON architecture consists of 'k' layers of array processors. The lowest layers consist of lower-level processing layers, and the rest consist of higher-level processing layers. The interconnectivity of the PEs in each array is based on a full hexagonal mesh structure. KYDON uses graph models to represent and process the knowledge, extracted from the image. The knowledge base of KYDON is distributed among its PE's.

This paper deals with the representation and acquisition of knowledge extracted during a real-time path planning process in an unknown space. In particular, the extracted knowledge from each current free navigation space is represented by attributed graphs. The graph knowledge representation forms are “processed and combined” appropriately by generating a scheme which acquires the extracted knowledge during the navigation. Illustrated examples are provided for simple navigation cases.

An automatic visual inspection-diagnosis expert system based on knowledge extraction and acquisition is presented. Images of a PC-board are grabbed and fed to the system. The main goal of the system is to acquire the functional behavior of the PC-board and compare the results with experimental ones. The sytem can also respond to other user queries. The secondary goal would be to detect any type of physical or behavioral defects that may be present on the board.

A nonconventional text character extraction technique, called horizontal-vertical projection (HVP), has been used for feature extraction of numerous alphanumeric symbols. In an earlier paper a classification algorithm based on a 1-D Fourier transform (FT) was used with NVP. An alternative 2-D FT feature extraction technique is presented and shown to improve on the 1-D approach.>

In this panel session, the following topics are discussed: artificial intelligence in business; artificial intelligence in multimedia; neural networks as a tool for artificial intelligence: software engineering for knowledge-based systems: and artificial intelligence as a solution for software engineering.>

Deals with the specifications for the development of an expert tool for the automatic optical detection and recognition of the connectivity among devices in digital electronic circuits and the understanding of the circuits functionality. In particular, the proposed tool, called ANTISTROFEAS, uses expert knowledge recognizing and understanding electronic circuits without the use of their associated database. The ANTISTROFEAS tool will use classical picture processing methods in combination with heuristics and knowledge acquisition schemes. The expert tool proposed is used for understanding of 'unknown' electronic circuits, or where the complexity of the circuit is too great, so that any human searching effort requires long time for reliable results.>

The efficient recognition of typed characters is one of the interesting topics of research with significant application merit in the computerized administration and library domains. In this paper, we present the results of a non-conventional text character extraction technique, called the Horizontal–Vertical Projections (HVP) method. The development of a knowledge base by using Fourier Transform as a tool which creates the characters' main features for the knowledge base, improves the performance of the HVP as a text character recognition technique. Various problems and their solutions are discussed with the presentation of the knowledge-based character recognition process.

This paper elaborates on the data modeling and data sharing issues that arise when developing and interconnecting modern software development environments. It primarily focuses on the investigation of adjoined object management issues for tool integration, software development environment extensibility, and interconnectivity.

A panel session concerning the use of knowledge-based techniques is reported. The discussion covers the following topics: knowledge-based machine vision architectures; experiences from a knowledge-based integrated-circuit testing system; artificial-intelligence (AI) tools for requirements capture and analysis; an approach to knowledge modeling in expert systems design; and a homogeneous approach to bringing AI and database management systems together.>

The coding of a reduced amount of picture information for a high speed transmission is discussed. The authors present a recursive algorithm ZETA, which scans the elements of a 2-D picture P(N,N) implementing the basis of a hierarchical pyramid. A simple average procedure is also provided, which accepts as input the N/sup 2/ pixels produced by the ZETA algorithm and constructs the exponential pyramid data structure of (4N/sup 2/-1)/3 elements for the picture P. This pyramidal scheme is used to produce the minimal picture tree by applying the regular decomposition methodology. A hierarchical orthogonal transformation is applied on the minimal picture tree realizing the reduced picture coding.>

This paper presents an efficient method for transmitting regularly-decomposed pictures which have been coded through an orthogonal Walsh-Hadamard transformation. The regular-decomposition procedure converts the picture into a tree representation and compresses the original amount of data by not retaining non-informative areas of the picture. The Walsh-Hadamard transformation is performed on the regularly-decomposed picture as an averaging or smearing process to allow transmission which is more immune to channel errors. It also makes bandwidth reduction possible because, since the image energy which is usually uniformly distributed in the spacial domain now tends to be concentrated near the origin of the transform domain, some of the low-magnitude samples may be discarded. A conceptual design is given where the three steps of regular-decomposition, Walsh-Hadamard transformation, and transmission, instead of being performed sequentially one after the other, are now carried out simultaneously.