Social unrest, a multifaceted phenomenon that is influenced by a variety of interconnected factors, presents substantial obstacles to societal stability and governance. The comprehension of local nuances is frequently restricted by the analysis of drivers of unrest at broad geographic scales or the isolation of specific causes in traditional studies. This paper introduces the SCEIGE framework, which classifies unrest drivers into six essential categories: Socio-demographic, Cultural, Environmental, Infrastructural, Geographic, and Economic. This framework is designed to address these challenges. SCEIGE offers a comprehensive perspective on the fundamental causes of social unrest by modeling geographic spaces at fine resolutions and incorporating a wide range of variables. We further enhance this framework by introducing a novel clustering and machine learning methodology, SC-XG (SCEIGE Clustering with XGBoost), which organizes geographic regions according to SCEIGE patterns. SC-XG not only reveals the local drivers of unrest but also facilitates the predictive analysis of social unrest events. This paper also illustrates the effectiveness of high-resolution SCEIGE geo-rasters in analyzing social unrest and confirms that the drivers of unrest differ across regions, underscoring the necessity of a local-level understanding. We identify critical, region-specific unrest drivers and address the broader implications for predicting and mitigating social unrest globally by applying SC-XG to unrest patterns in India.

Two key challenges lie in the facial attractiveness computation research: the lack of discriminative face representations, and the scarcity of sufficient and complete training data. Motivated by recent promising work in face recognition using deep neural networks to learn effective features, the first challenge is expected to be addressed from a deep learning point of view. A very deep residual network is utilized to enable automatic learning of hierarchical aesthetics representation. The inspiration to deal with the second challenge comes from the natural representation of the training data, where each training face can be associated with a label (score) distribution given by human raters rather than a single label (average score). This paper, therefore, recasts facial attractiveness computation as a label distribution learning problem. Integrating these two ideas, an end-to-end attractiveness learning framework is established. We also perform feature-level fusion by incorporating the low-level geometric features to further improve the computational performance. Extensive experiments are conducted on a standard benchmark, the SCUT-FBP dataset, where our approach shows significant advantages over the other state-of-the-art work.

Social Unrest Reconnaissance Gazetteer (or SURGE) is a Web-based application that provides an open system to visualize and integrate spatio-temporal data about social unrest events with related data layers in South Asia to facilitate data-driven as well as model-based investigations and analyses. Currently, the system displays eight categories of unrest, based primarily on the Global Database of Events, Language and Tone (GDELT) and the Global Terrorism Database (GTD). Users have the ability to select a single day or a range of dates along with the category of unrest they are interested to investigate. The users also have the option to normalize the raw event counts by population density. Additionally, the users can view infrastructure layers that facilitate or hinder the diffusion of unrest events (e.g., collated from an open GIS data-source: OpenStreetMap (www.openstreetmap.org)) and choropleth layers to display various socio-economic indicators (e.g., derived from global surveys and government census data such as the 2011 India census data (cenusindia.gov.in) and IPUMS Terra (data.terrapop.org)). Currently, SURGE displays unrest events for India, Pakistan and Bangladesh as heat map layers in multiple spatial resolutions. Challenges have involved geo-synchronization, data conversions, and displaying multiple layers of dense geospatial datasets. Future capabilities include automatic ingestion of raw data and standardizing levels of unrest using significant predictors.

Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.

The past few years have witnessed impressive progress on the research of salient object detection. Nevertheless , existing approaches still cannot perform satisfactorily in the case of complex scenes, particularly when the salient objects have non- uniform appearance or complicated shapes, and the background is complexly structured. One important reason for such limitations may be that these approaches commonly ignore the factor of perceptual grouping in saliency modeling. To address this issue, this paper presents a novel computational model for object -based visual saliency, which explicitly takes into consideration the connections between attention and perceptual grouping, and incorporates Gestalt grouping cues into saliency computation. Inspired by the sensory enhancement theory, we suggest a paradigm for object-based saliency modeling, that is, object-based saliency stems from spreading attention along Gestalt grouping cues. Computationally , three typical Gestalt cues, including proximity, similarity, and closure, are respectively extracted from the given image, which are then integrated by constructing a unified Gestalt graph. A new algorithm named personalized power iteration clustering is developed to effectively fulfill the spreading of attention information across the Gestalt graph. Intensive experiments have been carried out to demonstrate the superior performance of the proposed model in comparison to the state-of-the-art.

Silent speech interfaces (SSIs), which recognize speech from articulatory information (i.e., without using audio information), have the potential to enable persons with laryngectomy or a neurological disease to produce synthesized speech with a natural sounding voice using their tongue and lips. Current approaches to SSIs have largely relied on speaker-dependent recognition models to minimize the negative effects of talker variation on recognition accuracy. Speaker-independent approaches are needed to reduce the large amount of training data required from each user; only limited articulatory samples are often available for persons with moderate to severe speech impairments, due to the logistic difficulty of data collection. This paper reported an across-speaker articulatory normalization approach based on Procrustes matching, a bidimensional regression technique for removing translational, scaling, and rotational effects of spatial data. A dataset of short functional sentences was collected from seven English talkers. A support vector machine was then trained to classify sentences based on normalized tongue and lip movements. Speaker-independent classification accuracy (tested using leave-one-subject-out cross validation) improved significantly, from 68.63 % to 95.90%, following normalization. These results support the feasibility of a speaker-independent SSI using Procrustes matching as the basis for articulatory normalization across speakers. Index Terms: silent speech recognition, speech kinematics, Procrustes analysis, support vector machine

Speech sounds are the result of coordinated movements of individual articulators. Understanding each articulator's role in speech is fundamental not only for understanding how speech is produced, but also for optimizing speech assessments and treatments. In this paper, we studied the individual contributions of six articulators, tongue tip, tongue blade, tongue body front, tongue body back, upper lip, and lower lip to phoneme classification. A total of 3,838 vowel and consonant production samples were collected from eleven native English speakers. The results of speech movement classification using a support vector machine indicated that the tongue encoded significantly more information than lips, and that the tongue tip may be the most important single articulator among all of the six for phoneme production. Furthermore, our results suggested that the tracking of four articulators (i.e., tongue tip, tongue body back, upper lip, and lower lip) may be sufficient for distinguishing major English phonemes based on articulatory movements.

Recent research has demonstrated the potential of using an articulation-based silent speech interface for command-and-control systems. Such an interface converts articulation to words that can then drive a text-to-speech synthesizer. In this paper, we have proposed a novel near-time algorithm to recognize whole-sentences from continuous tongue and lip movements. Our goal is to assist persons who are aphonic or have a severe motor speech impairment to produce functional speech using their tongue and lips. Our algorithm was tested using a functional sentence data set collected from ten speakers (3012 utterances). The average accuracy was 94.89% with an average latency of 3.11 seconds for each sentence prediction. The results indicate the effectiveness of our approach and its potential for building a real-time articulation-based silent speech interface for clinical applications.

Articulation-based silent speech interfaces convert silently produced speech movements into audible words. These systems are still in their experimental stages, but have significant potential for facilitating oral communication in persons with laryngectomy or speech impairments. In this paper, we report the result of a novel, real-time algorithm that recognizes whole-words based on articulatory movements. This approach differs from prior work that has focused primarily on phoneme-level recognition based on articulatory features. On average, our algorithm missed 1.93 words in a sequence of twenty-five words with an average latency of 0.79 seconds for each word prediction using a data set of 5,500 isolated word samples collected from ten speakers. The results demonstrate the effectiveness of our approach and its potential for building a real-time articulation-based silent speech interface for health applications.

Redistricting is the process of dividing a geographic area consisting of spatial units-often represented as spatial polygons-into smaller districts that satisfy some properties. It can therefore be formulated as a set partitioning problem where the objective is to cluster the set of spatial polygons into groups such that a value function is maximized [1]. Widely used algorithms developed for point-based data sets are not readily applicable because polygons introduce the concepts of spatial contiguity and other topological properties that cannot be captured by representing polygons as points. Furthermore, when clustering polygons, constraints such as spatial contiguity and unit distributedness should be strategically addressed. Toward this, we have developed the Constrained Polygonal Spatial Clustering (CPSC) algorithm based on the A* search algorithm that integrates cluster-level and instance-level constraints as heuristic functions. Using these heuristics, CPSC identifies the initial seeds, determines the best cluster to grow, and selects the best polygon to be added to the best cluster. We have devised two extensions of CPSC-CPSC* and CPSC*-PS-for problems where constraints can be soft or relaxed. Finally, we compare our algorithm with graph partitioning, simulated annealing, and genetic algorithm-based approaches in two applications-congressional redistricting and school districting.

The articulatory distinctiveness among vowels has been frequently characterized descriptively based on tongue height and front-back position; however, very few empirical methods have been proposed to characterize vowels based on time-varying articulatory characteristics. Such information is not only needed to improve knowledge about the articulation of vowels but also to determine the contribution of articulatory imprecision to poor speech intelligibility. In this paper, a novel statistical shape analysis was used to derive a vowel space that depicted the quantified articulatory distinctiveness among vowels based on tongue and lip movements. The effectiveness of the approach was supported by vowel classification accuracy of up to 91.7%. The theoretical relevance and clinical implication of the derived vowel space were discussed. Index Terms: speech production, articulatory vowel space, Procrustes analysis, multi-dimensional scaling

Learning objects (LOs) have been previously used in computer science education. However, analyses in previous studies have been limited to surveys with limited numbers of LOs and students. The lack of copious quantitative data on how LOs impact student learning makes detailed analysis of LO usefulness problematic. Using an empirical approach, we have studied a suite of LOs, comprehensive in both the content covered and the range of difficulty, deployed to CS1 courses from 2007-2010. We review previous work on predictors of achievement and impact of active learning and feedback. We also provide a high-level overview of our LO deployment. Finally, based on our analysis of student interaction data, we found that (1) students using LOs have significantly higher assessment scores than the control group, (2) several student attributes are significant predictors of learning, (3) active learning has a significant effect on student assessment scores, and (4) feedback does not have a significant effect, but there are variables with significant moderating effects.

Learning objects (LO) have previously been used to help deliver introductory computer science (CS) courses to students. Students in such introductory CS courses have diverse backgrounds and characteristics requiring revision to LO content and assessment to promote learning in all students. However, revising LOs in an ad hoc manner could make student learning harder for subsequent deployments. To address this problem, we present a systematic revision process for LOs (LOSRP) using proven techniques from educational research including Bloom's Taxonomy levels, item-total correlation, and Cronbach's Alpha. LOSRP uses these validation methods to answer seven questions in order to diagnose what needs to be revised in the LO. Then, LOSRP provides guidelines on revising LOs for each of the seven questions. As an example, we discuss how LOSRP was used to revise the content and assessment for 16 LOs deployed to over 400 students in introductory CS courses in 2009. Lastly, although initially designed for LO revision, we briefly discuss how LOSRP could be used for assessment revision in intelligent tutoring systems.

We present a framework for the automatic annotation of learning objects (LOs) with empirical usage metadata. Our implementation of the Intelligent Learning Object Guide (iLOG) was used to collect interaction data of over 200 students' interactions with eight LOs. We show that iLOG successfully tracks student interaction data that can be used to automate the creation of meaningful empirical usage metadata that is based on real-world usage and student outcomes.

Clustering is an important task in spatial data mining and spatial analysis. We propose a clustering algorithm P-DBSCAN to cluster polygons in space. P-DBSCAN is based on the well established density-based clustering algorithm DBSCAN. In order to cluster polygons, we incorporate their topological and spatial properties in the process of clustering by using a distance function customized for the polygon space. The objective of our clustering algorithm is to produce spatially compact clusters. We measure the compactness of the clusters produced using P-DBSCAN and compare it with the clusters formed using DBSCAN, using the Schwartzberg index. We measure the effectiveness and robustness of our algorithm using a synthetic dataset and two real datasets. Results show that the clusters produced using P-DBSCAN have a lower compactness index (hence more compact) than DBSCAN.

The traditional point-based clustering algorithms when applied to geospatial polygons may produce clusters that are spatially disjoint due to their inability to consider various types of spatial relationships between polygons. In this paper, we propose to represent geospatial polygons as sets of spatial and non-spatial attributes. By representing a polygon as a set of spatial and non-spatial attributes we are able to take into account all the properties of a polygon (such as structural, topological and directional) that were ignored while using point-based representation of polygons, and that aid in the formation of high quality clusters. Based on this framework we propose a dissimilarity function that can be plugged into common state-of-the-art spatial clustering algorithms. The result is clusters of polygons that are more compact in terms of cluster validity and spatial contiguity. We show the effectiveness and robustness of our approach by applying our dissimilarity function on the traditional k-means clustering algorithm and testing it on a watershed dataset.

Redistricting is the process of dividing a geographic area into districts or zones. This process has been considered in the past as a problem that is computationally too complex for an automated system to be developed that can produce unbiased plans. In this paper we present a novel method for redistricting a geographic area using a heuristic-based approach for polygonal spatial clustering. While clustering geospatial polygons several complex issues need to be addressed - such as: removing order dependency, clustering all polygons assuming no outliers, and strategically utilizing domain knowledge to guide the clustering process. In order to address these special needs, we have developed the constrained polygonal spatial clustering (CPSC) algorithm that holistically integrates do-main knowledge in the form of cluster-level and instance-level constraints and uses heuristic functions to grow clusters. In order to illustrate the usefulness of our algorithm we have applied it to the problem of formation of unbiased congressional districts. Furthermore, we compare and contrast our algorithm with two other approaches proposed in the literature for redistricting, namely-graph partitioning and simulated annealing.

We report on a recently funded project called Renaissance Computing, an initiative for promoting student participation in computing. We propose a radical re-thinking not only of our core curriculum in CS, but of the role of CS at the university level. In our conception, ''computational thinking'' is neither easily separated from other endeavors nor easily balkanized into a single department. We thus imagine a CS curriculum that is inextricably linked to other domains. Our proposed initiative covers introductory, depth, and capstone courses, targeting both CS majors and minors. It is also aimed to develop interdisciplinary CS courses in sciences, engineering, arts, and humanities. Furthermore, the framework embraces collaborative learning to help improve learning.

Widespread use of GPS devices and explosion of remotely sensed geospatial images along with cheap storage devices has resulted in vast amounts of data. More recently, with the advent of wireless technology, a large number of sensor networks have been deployed to monitor many human, biological and natural processes. This poses a challenge in many data rich application domains. The problem now is how best to choose the datasets to solve specific problems. Some of the datasets may be redundant and their inclusion in analysis may not only be time consuming, but may lead to erroneous conclusions. We propose the concept of data support as the basis for efficient, cost-effective and intelligent use of geospatial data in order to reduce uncertainty in the analysis and consequently in the results. Data support is defined as the process of determining the information utility of a data source to help decide which one to include or exclude to improve cost-effectiveness in existing data analysis. In this article we use mutual information as the basis of computing data support. The concept of mutual information is defined in information theory as a measure to compute information gain or loss between two disjoint datasets. We use this to compute the optimal datasets in specific applications. The effectiveness of the approach is demonstrated using an application in the hydrological analysis domain.

Extraction of features from images has been a goal of researchers since the early days of remote sensing. While significant progress has been made in several applications, much remains to be done in the area of accurate identification of high-level features such as buildings and roads. This paper presents an approach for detecting bridges over water bodies from multispectral imagery. The multispectral image is first classified into eight land-cover types using a majority-must-be-granted logic based on the multiseed supervised classification technique. The classified image is then categorized into a trilevel image: water, concrete, and background. Bridges are then recognized in this trilevel image by using a knowledge-based approach that exploits the spatial arrangement of bridges and their surroundings using a five-step approach. A river extraction module identifies the rivers using a recursive scanning technique and geometric constraints. Using a neighborhood operator and the knowledge of the spatial dimensions of a typical bridge, we identify the possible bridge pixels. These potential bridge pixels are then grouped into possible bridge segments based on their connectivity and geometric properties. Finally, these bridge segments are verified on the basis of directional water index along different directions and their connectivity with the road segments. The approach proposed in this paper has been implemented and tested with images from the IRS-1C/1-D satellite that has a spatial resolution of 23.5 23.5 m. The results show that this approach is both efficient and effective in extracting bridges.

A new method for estimating fractal dimension of tree crowns from digital images is presented. Three species of trees, Japanese yew (Taxus cuspidata Sieb & Zucc), Hicks yew (Taxus × media), and eastern white pine (Pinus strobus L.), were studied. Fractal dimensions of Japanese yew and Hicks yew range from 2.26 to 2.70. Fractal dimension of eastern white pine range from 2.14 to 2.43. The difference in fractal dimension between Japanese yew and eastern white pine was statistically significant at 0.05 significance level as was the difference in fractal dimension between Hicks yew and eastern white pine. On average, the greater fractal dimensions of Japanese yew and Hicks yew were possibly related to uniform foliage distribution within their tree crowns. Therefore, fractal dimension may be useful for tree crown structure classification and for indexing tree images.

This paper studies how biologically meaningful landmarks extracted from face images can be exploited for face recognition using the bidimensional regression. Incorporating the correlation statistics of landmarks, this paper also proposes a new approach called eigenvalue weighted bidimensional regression. Complex principal component analysis is used for computing eigenvalues and removing correlation among landmarks. We evaluate our approach using two standard face databases: the Purdue AR and the NIST FERET. Experimental results show that the bidimensional regression is an efficient method to exploit geometry information of face images.

A learning object is a structured, standalone media resource that encapsulates high quality information to facilitate learning and pedagogy. In this paper, we describe our approach to design, develop, and validate learning objects for CS1. In particular, we focus on one learning object that teaches students about classes and objects. SCORM (Shareable Content Object Reference Model) standards and ACM/IEEE-CS Computing Curriculum 2001 form the basis of our design. Each learning object is self-contained and by design, the length of the content section is kept short to retain student interest. The learning object has a glossary providing definitions to key terms and a help menu. Each learning object covers a core Computer Science topic addressed by four components: (1) A brief tutorial or explanation including definitions, rules, and principles, (2) A set of real-world examples illustrates key concepts and includes worked examples and problems, models, and sample code, (3) A set of practice exercises provides important active experiences to the student, with constructive feedback to student responses, (4) A set of problems graded by the computer provides a final assessment. Our instructional design also incorporates theories of multimedia learning, providing guidance on the effective combination of text, graphics audio, and Flash animation. We also report on a pilot evaluation where students rated the learning object highly in terms of its design, usefulness, and appropriateness. We present student achievement results, comparing achievement of students participating in traditional face-to-face laboratory activities versus students using the Web-based learning object. A between-group post-test only research design showed no significant achievement difference between the two groups. Results confirm our belief that the use of modular, Web-based learning objects can be used successfully for independent learning and are a viable option for distance delivery of course components. Encouraged by these results, our project and research is continuing Fall 2004, with the development of additional learning objects and instrumentation mechanisms tracking real-time dynamic activity-based data.The "Practice Exercises" section of our "Simple Class" learning object, for example, has four exercise modules: (1) class identification, where students are asked to identify whether an item is an appropriate candidate as a class (Abraham Lincoln vs. President, for example), (2) data members and methods, where students interact with an animation (with sound) to identify the appropriate data members for a dog class, (3) dissect a class definition, where students are given code with highlighted segments and are asked to label each segment into either "class", "method name", "data member", or "method body", and (4) building a class, where students are given a heterogeneous set of data members and methods, and must pick the appropriate ones to build a class; if the selection is correct, the Java-based class will be expanded accordingly with specific Java code. For each exercise, we provide extensive real-time feedback for each response. Figure 1 shows a screen shot of one of the exercises on data members and methods.

Closed laboratories are becoming an increasingly popular approach to teaching introductory computer science courses. However, as observed in [1], “Considering the prevalence of closed labs and the fact that they have been in place in CS curricula for more than a decade, there is little published evidence assessing their effectiveness. ” In this paper, we report on how students’ performance in closed laboratories relates to their performances on a placement exam, homework assignments, course exams, and how it relates to their self-reported attitudes towards our CS1 course. This analysis provides insights to help us improve the design of our laboratories as well as other components of CS1.

An introductory CS1 course presents problems for educators and students due to students' diverse background in programming knowledge and exposure. Students who enroll in CS1 also have different expectations and motivations. Prompted by the curricular guidelines for undergraduate programs in computer science released in 2001 by the ACM/IEEE, and driven by a departmental project to reinvent the undergraduate computer science and computer engineering curricula at the University of Nebraska-Lincoln, we are currently implementing a series of changes which will improve our introductory courses. One key component of our project is an online placement examination tied to the cognitive domain that assesses student knowledge and intellectual skills. Our placement test is also integrated into a comprehensive educational research design containing a pre- and post-test framework for assessing student learning. In this paper, we focus on the design and implementation of our placement exam and present an analysis of the data collected to date.

Closed laboratories are becoming an increasingly popular approach to teaching introductory computer science courses, as they facilitate structured problem-solving and cooperation. However, most closed laboratories have been designed and implemented without embedded instructional research components for constant evaluation of the laboratories' effectiveness. As a result, it is not convenient to maintain and improve the laboratories over time so that they adapt to changing CS topics, curricula, and student needs. This article reports on an integrated framework for designing, implementing, and maintaining laboratories with embedded instructional research design. Although the activities reported here are part of our department-wide effort to cover CS0, CS1, and CS2, we focus here on the design and implementation of the labs for CS1.

A Geographic Information System (GIS) populated with disparate data sources has multiple and different representations of the same real-world object. Often, the type of information in these sources is different, and combining them to generate one composite representation has many benefits. The first step in this conflation process is to identify the features in different sources that represent the same real-world entity. The matching process is not simple, since the identified features from different sources do not always match in their location, extent, and description. We present a new approach to matching GIS features from disparate sources. A graph theoretic approach is used to model the geographic context and to determine the matching features from multiple sources. Experiments on implementation of this approach demonstrate its viability.

Secure one-to-many multicasting has been a popular research area in the past. Secure many-to-many multicasting is becoming popular with applications such as private conferencing and distributed interactive simulation. Most of the existing secure multicasting protocols use a centralized group manager to enforce access control and for key distribution. In the presence of multiple senders it is desirable to delegate group management responsibility to all the senders. We propose a distributed group key management scheme to support secure many-to-many communication. We divide key distribution overhead evenly among the senders. Our protocol is scalable and places equal trust in all the senders.

The black layer is digitized from a USGS topographic map digitized at 1000 dpi. The connected components of this layer are analyzed and separated into line art, text, and icons in two passes. The paired street casings are converted to polylines by vectorization and associated with street labels from the character recognition phase. The accuracy of character recognition is shown to improve by taking account of the frequently occurring overlap of line art with street labels. The experiments show that complete vectorization of the black line-layer bitmap is the major remaining problem.

We propose a dual encryption protocol for scalable secure multicasting. Multicasting is a scalable solution for group communication. It however poses several unique security problems. We use hierarchical subgrouping to achieve scalability. Third-party hosts or members of the multicast group are designated as subgroup managers. They are responsible for secret key distribution and group membership management at the subgroup level. Unlike existing secure multicast protocols, our protocol need not trust the subgroup managers with the distribution of data encryption keys. The dual encryption protocol proposed in this paper distributes encrypted data encryption keys via subgroup managers. We also present a classification of the existing secure multicast protocols, compare their relative merits and show the advantages of our protocol.

We present a system for recognizing a large class of engineering drawings characterized by alternating instances of symbols and connection lines. The class includes domains such as flowcharts, logic and electrical circuits, and chemical plant diagrams. The output of the system, a netlist identifying the symbol types and interconnections, may be used for design simulation or as a compact portable representation of the drawing. The automatic recognition task is divided into two stages: 1) domain-independent rules are used to segment symbols from connection lines in the drawing image that has been thinned, vectorized, and preprocessed in routine ways; 2) a drawing understanding subsystem works in concert with a set of domain-specific matchers to classify symbols and correct errors automatically. A graphical user interface is provided to correct residual errors interactively and to log data for reporting errors objectively. The system has been tested on a database of 64 printed images drawn from text books and handbooks in different domains and scanned at 150 and 300 dpi resolution.

A genetic algorithm (GA) is a robust problem-solving method based on natural selection. Hardware's speed advantage and its ability to parallelize offer great rewards to genetic algorithms. Speedups of 1-3 orders of magnitude have been observed when frequently used software routines were implemented in hardware by way of reprogrammable field-programmable gate arrays (FPGAs). Reprogrammability is essential in a general-purpose GA engine because certain GA modules require changeability (e.g. the function to be optimized by the GA). Thus a hardware-based GA is both feasible and desirable. A fully functional hardware-based genetic algorithm (the HGA) is presented here as a proof-of-concept system. It was designed using VHDL to allow for easy scalability. It is designed to act as a coprocessor with the CPU of a PC. The user programs the FPGAs which implement the function to be optimized. Other GA parameters may also be specified by the user. Simulation results and performance analyses of the HGA are presented. A prototype HGA is described and compared to a similar GA implemented in software. In the simple tests, the prototype took about 6% as many clock cycles to run as the software-based GA. Further suggested improvements could realistically make the HGA 2–3 orders of magnitude faster than the software-based GA.

We present a complete system for recognizing a large class of symbolic engineering drawings that includes flowcharts, chemical plant diagrams, and logic & electrical circuits. The output of the system, a netlist identifying the symbol types and interconnections, may be used for design verification or as a compact portable representation of the drawing. The automatic recognition task is done in two stages: (1) domain-independent rules segment symbols from connection lines in the preprocessed drawing image and (2) an understanding subsystem makes use of a set of domain-specific matchers to classify symbols and correct errors automatically. A graphical user interface is provided to correct residual errors interactively. The system has been tested on a large database of printed images drawn from four different domains.

Face detection is integral to any automatic face recognition system. The goal of this research is to develop a system that performs the task of human face detection automatically in a scene. A system to correctly locate and identify human faces will find several applications, some examples are criminal identification and authentication in secure systems. This work presents a new approach based on principal component analysis. Face silhouettes instead of intensity images are used for this research. It results in reduction in both space and processing time. A set of basis face silhouettes are obtained using principal component analysis. These are then used with a Hough-like technique to detect faces. The results show that the approach is robust, accurate and reasonably fast.

Abstract Computer vision, owing to the size and complexity of its tasks and its importance to industrial and economic growth, was selected as one of the grand challenge problems by the U.S. Federal High Performance Computing Program. Integration of vision operations is identified as a key element of the challenge. A system to integrate computer vision in a distributed environment is presented here. This system, called DeViouS, is based on the client/server model and runs in a heterogeneous environment of Unix workstations. Modern computing environments include large numbers of high‐powered workstations connected by a very fast network. Many of these computers are idle most of the time. DeViouS takes advantage of this feature of computing environments to distribute the execution of vision tasks. Two primary goals of DeViouS are to provide a practical distributed system and a research environment for vision computing. DeViouS is based on a modular design that allows experimentation in various aspects of algorithm design, scheduling and network programming. It can make use of any existing computer vision packages with very minor changes to DeViouS. DeViouS has been tested in an environment of SUN and Digital workstations and has shown substantial improvements in speed over sequential computing with negligible overhead.

We present a system for the integration of computer vision tasks in a distributed environment. This system, called DeViouS, is based on the client/server model and runs in a heterogeneous environment of Unix workstations. It takes advantage of the free cycles in modern workstation environments to distribute and speed up the execution of vision tasks. Two primary goals of DeViouS are to provide a practical distributed system and a research environment for vision computing. DeViouS is based on a modular design that allows experimentation in various aspects of algorithm design, scheduling and network programming. It can make use of any existing computer vision package with very minor changes to DeViouS. DeViouS has been tested in an environment of SUN and Digital workstations and has shown substantial improvements in speed over sequential computing with negligible overhead.>

Recognizing naturally occurring objects has been a difficult task in computer vision. One of the keys to recognizing objects is the development of a suitable model. One type of model, the fractal, has been used successfully to model complex natural objects. A class of fractals, the L-system, has not only been used to model natural plants, but has also aided in their recognition. This research extends the work in plant recognition using L-systems in two ways. Stochastic L-systems are used to model and generate more realistic plants. Furthermore, to handle the complexity of recognition, a learning system is used that automatically generates a decision tree for classification. Results indicate that the approach used here has great potential as a method for recognition of natural objects.>

The main focus in parallel computer vision has so far been the design and analysis of parallel algorithms to perform individual operations. While it clearly is necessary and useful, it is not the ultimate goal. The goal is to design and implement a parallel vision system which integrates all the parallel vision modules easily and efficiently. A framework for building a parallel vision system is presented. The necessary and desirable features of such a system are identified. An initial design which incorporates these features is given.>

The split-level relaxation technique is analyzed in a parallel processing framework. It is shown that there is much parallelism inherent in the algorithm that can be exploited. This has been confirmed by implementation of the algorithm on an actual multiprocessor. Although the results are good, the implementation can be made more efficient. The use of multiple queues instead of a centralized queue can reduce memory contention, particularly in large multiprocessors. An asynchronous implementation can also improve the performance.>

The goal of high level vision is to identify a set of regions in a given image. This has been called by various names: the scene labeling problem’, the consistent labeling problem2, the constraint satisfaction problem3, Waltz filtering4, the satisfying assignment problem5, etc. There are several approaches to solve this problem, including backtracking, graph matching and relaxation. A new method called split-level relaxation, which is based on discrete relaxation was proposed in Ref. 6. It takes care of multiple semantic constraints by considering each of them independently. The problem is known to be NP-complete, so it takes a long time to solve. With the advent of multiprocessors, it is now imperative to see if the problem can be solved faster in the average case. In this paper we give a framework for solving the scene analysis problem in a parallel processing environment, using split-level relaxation. Experiments done on a multiprocessor show that it is indeed advantageous to use multiprocessors to solve this problem.