There has recently been considerable interest in automatic detection strategies for recognising application layer security threats such as Hypertext Transfer Protocol (HTTP) Slow Denial of-Service (Slow DoS) attacks in Internet of Things (IoT) networks. Most existing approaches however, fail to take cognisance of the substantial resource constraints imposed upon IoT environments, which limits the applicability and deployment of many Slow DoS detection mechanisms. This paper addresses this significant security threat for resource scarce IoT nodes and networks in proposing an accurate and computationally efficient approach to packet-based intrusion detection of HTTP Slow DoS activity. The paper both critically analyses and measures the impact of applying network attribute filtering and packet sampling to reduce the computational overheads on the resource constrained IoT Slow DoS detection node. The unique solution proposed uses a dataset synthesised from a live IoT environment comprising both legitimate and malicious network events in the form of legitimate HTTP traffic and Slow DoS attacks. Experimental results corroborate that combining filtering at the Border Router of only in-bound packets containing no TCP payload with a systematic packet sampling scheme at a sampling ratio of up to 1:64, the processing overheads on the detection node are significantly reduced. The novel contribution presented is a resource efficient solution, garnered by employing systematic sampling to seamlessly and accurately support selective attribute-based intrusion detection of HTTP Slow DoS attacks in IoT networks.

Medicinal plants are a primary source of disease treatment in many countries. As most are edible however, consumption of the wrong herbal plants can have serious consequences and even lead to death. Automatic accurate recognition of plant species to help users who do not have specialist knowledge of herbal plants is thus a desirable aim. Several automatic medicinal plant identification systems have been proposed, though most are significantly constrained either in the small number of species or in requiring manual image segmentation of plant leaves. This means they are captured on a plain background rather than being readily identified in their natural surroundings, which often involve complex and noisy backgrounds. While deep learning (DL) based methods have made considerable strides in recent times, their potential has not always been maximised because they are trained with samples which are not always fully representative of the intra-class and inter-class differences between the plant species concerned. This paper addresses this challenge by incorporating mutual information into a Convolutional Neural Network (CNN) model to select samples for the training, validation, and testing sets based on a similarity measure. A critical comparative evaluation of this new CNN medicinal plant classification model incorporating a mutual information guided training (MIGT) algorithm for sample selection, corroborates the superior classification performance achieved for the VNPlant-200 dataset, with an average accuracy of more than 97%, while the precision and recall values are also consistently above 97%. This is significantly better than existing CNN classification methods for this dataset as it crucially means false positive rates are substantially lower thus affording improved identification reliability.

Slow DoS attacks have proven to pose a significant security threat to low-resource IoT devices and networks, because they can be launched by nodes which consume nominal bandwidth and have limited resource capability. This makes such malicious attacks easy to initiate, but difficult to mitigate. There also exists the recurrent likelihood of misclassifying legitimate nodes, which are incurring slow or poor network connectivity, as malicious activity. Existing intrusion detection systems (IDS) for detecting Slow DoS attacks often require the creation of large datasets for post event analysis. A functional disadvantage of this dataset-driven approach is the sheer volume of data required, due to the high number of network attributes and events collated, which precludes an in-line, real-time IDS detection solution for live IoT networks. This paper presents an innovative IDS detection framework for resource constrained IoT networks. Using a set of only four attributes, a two-step analysis of live IoT network events enables Slow DoS attacks, in the form of Slowloris, to be both efficiently and reliably detected in real-time. In addition, this lightweight IDS framework can accurately distinguish between malicious and genuine nodes encountering slow or intermittent network connections.

A major challenge in spectrum sensing for cognitive radio (CR) applications is the very high sampling rates involved, which imposes significant demands on the signal acquisition technology. This has given impetus to applying compressive sensing (CS) as a sub‐Nyquist sampling paradigm for CR‐type wireless signals which exhibit sparsity in certain domains. CS architectures like the random demodulator (RD) and compressive multiplexer (CM) can be used for CR spectral sensing, though both are inherently restricted in terms of the signal classes they can effectively process. To address these limitations, this study presents two unified RD and CM‐based CS architectures that seamlessly integrate precolouring and the multitaper spectral estimator into their respective structures to facilitate efficient sensing of both digitally modulated and narrowband signals, along with popular CR‐access technologies like orthogonal frequency division multiplexing . A significant feature of these unified CS architectures is they do not require a priori knowledge of either the input signal or modulation scheme, while a tristate spectral classifier is introduced to afford notably enhanced spectrum access opportunities for unlicensed secondary users. A critical performance evaluation corroborates that both unified architectures demonstrate consistently superior CS results and robustness across a broad range of CR‐type signals, modulations and access technologies.

As mobile applications and services have developed, the dramatic growth in user data traffic has led to the legacy channels becoming ever more congested with the commensurate requirement for more spectrum. This has motivated both regulatory bodies and industry to investigate innovative strategies to increase the existing spectral efficiency. Prominent examples include both Long Term Evolution (LTE) which employs orthogonal frequency‐division modulation technology to improve bandwidth efficiency, and heterogeneous networks, which facilitate the offloading of data traffic between technologies such as from LTE to Wi‐Fi and vice versa. Furthermore, as 5G mobile technology and related standards mature, there is an impetus to address the issue of secondary user (SU) spectrum access in which TV White Space (TVWS) is the prime contender. Two nascent viewpoints have emerged as to how this will evolve: (i) greater coverage, (ii) increased throughput allied with lower latency. This study presents a novel TVWS framework that successfully fulfils both criteria to ensure 5G services can both exploit TVWS spectrum and protect the benefits of SU access and quality‐of‐service provision by using a routing strategy on the Access Network Discovery and Selection Function server to dynamically determine the most suitable heterogeneous technology for the new framework.

As more user applications emerge for wireless devices, the corresponding amount of traffic is rapidly expanding, with the corollary that ever‐greater spectrum capacity is required. Service providers are experiencing deployment blockages due to insufficient bandwidth being available to accommodate such devices. TV White Space (TVWS) represents an opportunity to supplement existing licensed spectrum by exploiting unlicensed resources. TVWS spectrum has materialised from the unused TV channels in the switchover from analogue to digital platforms. The main obstacles to TVWS adoption are reliable detection of primary users (PUs) i.e. TV operators and consumers, allied with specifically, the hidden node problem. This study presents a new Generalised Enhanced Detection Algorithm (GEDA) that exploits the unique way Digital Terrestrial TV (DTT) channels are deployed in different geographical areas. GEDA effectively transforms an energy detector into a feature sensor to achieve significant improvements in detection probability of a DTT PU. Furthermore, by framing a novel margin strategy utilising a keep‐out contour , the hidden node issue is resolved and a viable secondary user sensing solution formulated. Experimental results for a cognitive radio TVWS model have formalised both the bandwidth and throughput gains secured by TVWS users with this new paradigm.

Recently, base station (BS) sleeping has emerged as a viable conservation strategy for energy efficient communication networks. Switching-off particular BS during low-traffic periods requires the load to be sufficiently low so user performance is not compromised. There remain however, network energy saving opportunities during medium-to-high traffic periods if BSs operate in scalable fashion, which involves deploying multiple BSs with different power modes, i.e., macro/microcells, which are co-located in each cell. In this paper, a new scalable multimode BS switching (MMBS) cellular model is presented where depending on the traffic load, each BS operates in multimode: active, low-power and sleep, so the model dimensions network capacity by dynamically switching modes to minimise energy consumption. Results corroborate that the MMBS model reduces energy consumption by more than 50% during low-traffic and up to 9% during high-traffic conditions, thereby significantly improving the energy efficiency compared with the always-on and existing BS sleeping approaches.

While retinal images (RI) assist in the diagnosis of various eye conditions and diseases such as glaucoma and diabetic retinopathy, their innate features including low contrast homogeneous and non-uniformly illuminated regions, present a particular challenge for retinal image registration (RIR). Recently, the hybrid similarity measure, Expectation Maximization for Principal Component Analysis with Mutual Information (EMPCA-MI) has been proposed for RIR. This paper investigates incorporating various fixed and adaptive bin size selection strategies to estimate the probability distribution in the mutual information (MI) stage of EMPCA-MI, and analyses their corresponding effect upon RIR performance. Experimental results using a clinical mono-modal RI dataset confirms that adaptive bin size selection consistently provides both lower RIR errors and superior robustness compared to the empirically determined fixed bin sizes.

Images having either the same or different modalities can be aligned using the systematic process of image registration. Inherent image characteristics including intensity non-uniformities in magnetic resonance images and large homogeneous non-vascular regions in retinal and other generic image types however, pose a significant challenge to their registration. This paper presents an adaptive expectation maximisation for principal component analysis with mutual information (aEMPCA-MI) similarity measure for image registration. It introduces a novel iterative process to adaptively select the most significant principal components using Kaiser rule and applies 4-pixel connectivity for feature extraction together with Wichard's bin size selection in calculating the MI. Both quantitative and qualitative results on a diverse range of image datasets, conclusively demonstrate the superior image registration performance of «EMPCA-MI compared with existing Mi-based similarity measures.

Transmitting texture and depth images of captured camera view(s) of a 3D scene enables a receiver to synthesize novel virtual viewpoint images via Depth-Image-Based Rendering (DIBR). However, a DIBR-synthesized image often contains disocclusion holes, which are spatial regions in the virtual view image that were occluded by foreground objects in the captured camera view(s). In this paper, we propose to complete these disocclusion holes by exploiting the self-similarity characteristic of natural images via nonlocal template-matching (TM). Specifically, we first define self-similarity as nonlocal recurrences of pixel patches within the same image across different scales-one characterization of self-similarity in a given image is the scale range in which these patch recurrences take place. Then, at encoder we segment an image into multiple depth layers using available per-pixel depth values, and characterize self-similarity in each layer with a scale range; scale ranges for all layers are transmitted as side information to the decoder. At decoder, disocclusion holes are completed via TM on a per-layer basis by searching for similar patches within the designated scale range. Experimental results show that our method improves the quality of rendered images over previous disocclusion hole-filling algorithms by up to 3.9dB in PSNR.

Retinal images (RI) are widely used to diagnose a variety of eye conditions and diseases such as myopia and diabetic retinopathy. They are inherently characterised by having nonuniform illumination and low-contrast homogeneous regions which represent a unique set of challenges for retinal image registration (RIR). This paper investigates using the expectation maximization for principal component analysis based mutual information (EMPCA-MI) algorithm in RIR. It combines spatial features with mutual information to efficiently achieve improved registration performance. Experimental results for mono-modal RI datasets verify that EMPCA-MI together with Powell-Brent optimization affords superior robustness in comparison with existing RIR methods, including the geometrical features method.

Multimodal retinal images (RI) are extensively used for analysing various eye diseases and conditions such as myopia and diabetic retinopathy. The incorporation of either two or more RI modalities provides complementary structure information in the presence of non-uniform illumination and low-contrast homogeneous regions. It also presents significant challenges for retinal image registration (RIR). This paper investigates how the Expectation Maximization for Principal Component Analysis with Mutual Information (EMPCA-MI) algorithm can effectively achieve multimodal RIR. This iterative hybrid-based similarity measure combines spatial features with mutual information to provide enhanced registration without recourse to either segmentation or feature extraction. Experimental results for clinical multimodal RI datasets comprising colour fundus and scanning laser ophthalmoscope images confirm EMPCA-MI is able to consistently afford superior numerical and qualitative registration performance compared with existing RIR techniques, such as the bifurcation structures method.

Incorporating spatial features with mutual information (MI) has demonstrated superior image registration performance compared with traditional MI-based methods, particularly in the presence of noise and intensity non-uniformities (INU). This paper presents a new efficient MI-based similarity measure which applies Expectation Maximisation for Principal Component Analysis (EMPCA-MI), to afford significantly lower computational complexity, while providing analogous image registration performance with other feature-based MI solutions. Experimental analysis corroborates both the improved robustness and faster runtimes of EMPCA-MI, for different test datasets containing both INU and noise artefacts.

Traditionally the sliding window (SW) has been employed in vertex-based operational rate distortion (ORD) optimal shape coding algorithms to ensure consistent distortion (quality) measurement and improve computational efficiency. It also regulates the memory requirements for an encoder design enabling regular, symmetrical hardware implementations. This paper presents a series of new enhancements to existing techniques for determining the best SW-length within a rate-distortion (RD) framework, and analyses the nexus between SW-length and storage for ORD hardware realizations. In addition, it presents an efficient bit-allocation strategy for managing multiple shapes together with a generalized adaptive SW scheme which integrates localized curvature information (cornerity) on contour points with a bi-directional spatial distance, to afford a superior and more pragmatic SW design compared with existing adaptive SW solutions which are based on only cornerity values. Experimental results consistently corroborate the effectiveness of these new strategies.

Femtocell access points are inexpensive, plug and play home base stations designed to extend radio coverage and increase capacity within indoor environments. Their inherent uncoordinated and overlaid deployment however, means existing radio resource management (RRM) techniques are often ineffectual. Recent advances in dynamic RRM have emphasised the need for more efficient resource management strategies. While centralised resource management offers improved coordination and operator control giving better interference management, it is not scalable for increasing nodes. Distributed management techniques in contrast, do afford scaled deployment, but at higher node densities incur performance degradation in both system throughput and link-quality because of poor coordination. The level of spectrum sharing mandated by macro-femto deployment also impacts on system throughput and is scenario dependant. This paper presents a new hybrid resource management algorithm( HRMA) for down-links in orthogonal frequency division multiple access-based systems, with the model analysed for a range of macro-femto deployment scenarios. HRMA employs a dynamic fractional frequency reuse scheme for macro-cell deployment with frequency reuse defined for femto users depending on their location by making certain frequencies locally available based on macro-femto tier information sharing and efficient localised spectrum utilisation. Quantitative performance results confirm the efficacy of the HRMA strategy for various key system metrics including interference minimisation, outage probability and throughput.

Next generation heterogeneous wireless networks offer the end users with assurance of QoS inside each access network as well as during vertical handoff between them. For guaranteed QoS, the vertical handoff algorithm must be QoS aware, which cannot be achieved with the use of traditional RSS as the vertical handoff criteria. In this paper, we propose a novel vertical handoff algorithm which uses received SINR from various access networks as the handoff criteria. This algorithm consider the combined effects of SINR from different access networks with SINR value from one network being converted to equivalent SINR value to the target network, so the handoff algorithm can have the knowledge of achievable bandwidths from both access networks to make handoff decisions with QoS consideration. Analytical results confirm that the new SINR based vertical handoff algorithm can consistently offer the end user with maximum available bandwidth during vertical handoff contrary to the RSS based vertical handoff, whose performance differs under different network conditions. System level simulations also reveal the improvement of overall system throughputs using SINR based vertical handoff, comparing with the RSS based vertical handoff.

Since shape is regarded as one of the most important attributes of visualisation, it plays a pivotal role in semantic video object segmentation applications. One of the major objectives for the research community is to segment specific objects of interest from a video sequence using prescribed shape descriptors in a diverse range of applications from video surveillance and object tracking through to medical imaging. This paper addresses this challenge by presenting a new probabilistic spatio-temporal (PST) video object segmentation algorithm that incorporates a priori generic shape descriptor representations of particular objects in a sequence. The algorithm provides considerable improvement in perceptual picture quality compared with the existing PST segmentation technique, with the numerical analysis corroborating the superior subjective segmentation performance achieved.

This paper proposes a cross-layer protocol for energy-aware routing in wireless sensor networks. The protocol combines the energy depreciation rate, node distance and neighbourhood information from physical layer together with TDMA schedules from the MAC layer and also network life requirements from the application layer, to effectively determine the most efficient routes to the base station. This cross layer efficiency measure is then used to form dynamic clusters that adapt to changing traffic and energy conditions so assisting to both control the transmission power and schedule the sleep-wake cycles of nodes for better energy utilisation. The proposed protocol is a recursive aggregation scheme that transforms a network-wide routing dissemination problem into a single-hop query protocol that aids nodes in making multi-hop routing decisions based solely upon the information provided by their single-hop neighbours. Results confirm that the proposed technique balances the load on forwarding nodes, adapts the MAC layer precisely to the routing layer and minimizes data delivery time for increased traffic and large scale networks.

Embedding textural information into the probabilistic spatio-temporal (PST) video object segmentation is very important for achieving better segmentation, since this is one of the key perceptual attributes of any object. Existing video segmentation techniques however, ignore this feature because of the underlying difficulty in defining and hence characterizing a texture, which theoretically limits their segmentation performance. To address this problem, this paper proposes a new video object segmentation algorithm that involves a strategy to seamlessly incorporate texture information as a pixel feature in the PST framework. Experimental results for a variety of standard test sequences reveal a significant performance improvement in the quality of the video object segmentation achieved in comparison with the original PST method.

In the next generation multimedia wireless network environment that consists of heterogeneous access technologies, we need to offer the end user with multimedia QoS inside each access network as well as during vertical handoff between them. The vertical handoff algorithm have to be QoS aware, which cannot be achieved by using the traditional RSS as the vertical handoff criteria. In this paper, we propose a new vertical handoff algorithm using the receiving SINR from various access networks as the handoff criteria. By converting the different receiving SINR values, the handoff algorithm can have the knowledge of achievable bandwidths from both access networks, and make handoff decisions with multimedia QoS consideration. Analysis results confirms that the new SINR based vertical handoff algorithm is able to consistently offer the end user with maximum available bandwidth during vertical handoff comparing with the RSS based vertical handoff, whose performance differs under different network conditions.

Existing vertex-based operational rate-distortion (ORD) optimal shape coding algorithms can use a number of different distortion measurement techniques, including the shortest absolute distance (SAD), the distortion band (DB), the tolerance band (TB), and the accurate distortion measurement technique for shape coding (ADMSC). From a computational time perspective, an N-point contour requires O(N2) time for DB and TB for both polygon and B-spline-based encoding, while SAD and ADMSC incur O(N) time for polygonal encoding but O(N2) for B-spline based encoding, thereby rendering the ORD optimal algorithms computationally inefficient. This letter presents a novel distortion measurement strategy based on chord-length parameterization (DMCLP) of a boundary that incurs order O(N) complexity for both polygon and B-spline-based encoding while preserving a comparable rate-distortion performance to the original ORD optimal shape coding algorithms

Existing vertex-based operational rate-distortion (ORD) optimal shape coding algorithms use a vertex band around the shape boundary as the source of candidate control points (CP) usually in combination with a tolerance band (TB) and sliding window (SW) arrangement, as their distortion measuring technique. These algorithms however, employ a fixed vertex-band width irrespective of the shape and admissible distortion (AD), so the full bit-rate reduction potential is not fulfilled. Moreover, despite the causal impact of the SW-length upon both the bit-rate and computational-speed, there is no formal mechanism for determining the most suitable SW-length. This paper introduces the concept of a variable width admissible CP band and new adaptive SW-length selection strategy to address these issues. The presented quantitative and qualitative results analysis endorses the superior performance achieved by integrating these enhancements into the existing vertex-based ORD optimal algorithms.

Trading off computational complexity and quality is an important performance constraint for real time application of motion estimation algorithm. Previously, the novel concept of a distance-dependent thresholding search (DTS) was introduced for performance scalable motion estimation in video coding applications. This encompassed the full search as well as other fast searching techniques, such as the three-step search, with different threshold settings providing various quality-of-service levels in terms of processing speed and predicted image quality. The main drawback of the DTS was that the threshold values had to be manually defined. In this paper, the DTS algorithm has been extended to a fast and fully adaptive DTS (FADTS), a key feature of which is the automatic adaptation of the threshold using a desired target and the content from the actual video sequence, to achieve either a guaranteed level of quality or processing complexity. Experimental results confirm the performance of the FADTS algorithm in achieving this objective by demonstrating either comparable or improved search speed over existing fast algorithms including the diamond search, hexagon-based search, and enhanced hexagon-based search, while maintaining similar error performance

Any ad hoc ubiquitous network should be capable of taking proactive measures for smooth reconfiguration in the event of energy loss and a failure to key network Parent Nodes (PN). The overall objective of reconfiguration is to keep the network operating effectively for a prescribed time frame. The load profiles of PNs can be used to define its current state as well as to predict potential failures caused by energy loss due to high loads on particular PNs. This paper presents a new technique utilising historical load profile information about each PN to improve the estimate accuracy of its remaining lifetime before failure. A novel load balancing model is introduced that maintains PNs in a state whereby the time- to-live requirement is met by the PNs comprising the backbone of the network. Results prove that the proposed methodologies achieve the reconfiguration objectives as well as maintaining QoS, especially in high density ad hoc networks.

In this paper, a Guided Genetic Algorithm (GGA) has been presented for protein folding prediction (PFP) using 3D Hydrophobic-Hydrophilic (HP) model. Effective strategies have been formulated utilizing the core formation of the globular protein, which provides the guideline for the Genetic Algorithm (GA) while predicting protein folding. Building blocks containing Hydrophobic (H) -Hydrophilic (P or Polar) covalent bond are utilized such a way that it helps form a core that maximizes the fitness. A series of operators are developed including Diagonal Move and Tilt Move to assist in implementing the building blocks in three-dimensional space. The GGA outperformed Unger's GA in 3D HP model. The overall strategy incorporates a swing function that provides a mechanism to enable the GGA to test more potential solutions and also prevent it from developing a schema that may cause it to become trapped in local minima. Further, it helps the guidelines remain non-rigid. GGA provides improved and robust performance for PFP.

Traditional encodings for hydrophobic(H)-hydrophilic(P) model or HP lattice models is isomorphic, which adds unwanted variations for the same solution, thereby slowing convergence. In this paper a novel non-isomorphic encoding scheme is presented for HP lattice model, which constrains the search space. In addition, similarity comparisons are made easier and more consistent and it will be shown that non-deterministic search approach such as genetic algorithm (GA) converges faster when non-isomorphic encoding is employed

Most of gene regulatory network (GRN) studies are based on crisp and parametric algorithms, despite inherent fuzzy nature of gene co-regulation. This paper presents adaptive fuzzy evolutionary GRN Reconstruction (AFEGRN) framework for modeling GRNs. The AFEGRN automatically determines model parameters, such as, number of clusters for fuzzy c-means using fuzzy-PBM index and estimation of Gaussian distribution algorithm. The proposed strategy was tested for breast cancer and normal GRNs. The results conformed to biological knowledge and showed that most of cancer related GRN changes were caused by differentially expressed genes. This demonstrates effectiveness of AFEGRN to model any GRN.

From a video object segmentation perspective, using a joint spatio-temporal strategy is superior to processing with priority in either the spatial or temporal domains, as it considers a video sequence as a spatio-temporal grouping of pixels. However, existing spatio-temporal object segmentation techniques consider only pixel features, which tend to limit their performance in being able to segment arbitrary shaped objects. To address this limitation requires a new strategy for embedding generic shape information seamlessly into the segmentation process and this paper presents a new shape-based probabilistic spatio-temporal algorithm that achieves this objective. Experimental results using a number of standard video test sequences reveal a considerable performance improvement in being able to segment arbitrary shaped video objects in comparison with other contemporary space-time based video segmentation methods

Existing shape-based clustering algorithms, including fuzzy k-rings, fuzzy k-elliptical, circular c-shell, and fuzzy c-shell ellipsoidal are all designed to segment regular geometrically shaped objects such as circles, ellipses or combination of both. These algorithms however, are unsuitable for segmenting arbitrary-shaped objects, so in an attempt to address this issue, a fuzzy image segmentation of generic shaped clusters (FISG) algorithm was introduced that integrated generic shape information into the segmentation framework. It however, had a number of limitations relating to the mathematical derivation of the updated contour radius, the initial shape representation, and the impact of overlapping clusters. This paper proposes a new object based segmentation using fuzzy clustering (OSF) algorithm that solves these drawbacks by controlling the scaling of original shape, securing a better initial shape representation and avoids cluster overlapping, with both qualitative and quantitative results confirming the improved overall segmentation performance

Maximum likelihood (ML) is a popular and widely used statistical method, and while it is effective, its major short-comings are that it is a biased and non robust estimator. This paper proposes a formal establishment of an optimisation of ML (OML) by approximating the true distribution minimising the bias, and exploiting the underlying relationship between ML and the maximum entropy method. OML exposes the inefficiency of the classical ML in the orthogonal least square error minimisation sense, for a number of finite sample datasets. The robustness of the proposed OML method in finding an estimate within the boundaries of the parameter space is also proven. Under the same conditions, OML consistently provides a more global and efficient estimation, so both theoretically and empirically establishing its superiority over ML in terms of efficiency and robustness

Embedding generic shape information into probabilistic spatio-temporal video object segmentation is of pivotal importance to achieving better segmentation, since it provides valuable perceptual clues for humans in both distinguishing and recognising objects. Recently a probabilistic spatio-temporal video object segmentation algorithm incorporating shape information has been proposed, though since it is restricted to only pixel features, the probability of a pixel belonging to a certain cluster is directly correlated with its spatial location, which theoretically limits the segmentation performance of the technique. To address this problem, this paper proposes a new probabilistic spatio-temporal video object segmentation algorithm that incorporates generic shape information based on its region. Experimental results reveal a significant performance improvement in arbitrary-shaped video object segmentation compared with other contemporary methods for a variety of standard video test sequences.

Existing vertex-based operational-rate-distortion (ORD) optimal shape coding algorithms use a fixed width admissible control point band (FCB) around the shape boundary as the search space for possible control points. The width of the band however, is fixed and arbitrarily chosen independent of the admissible distortion and shape contour, so it fails to fully exploit the admissible control point band to reduce the bit-rate. This paper proposes a variable width admissible control point band (VCB) where the width associated to each boundary point is dynamically determined from the admissible peak distortion and shape information. In addition, the paper uses an accurate distortion measurement method to overcome a key limitation of existing distortion and tolerance band based methods. Experimental results reveal that both the qualitative and quantitative performance of the existing ORD algorithms are improved by seamlessly integrating the VCB and accurate distortion measuring approach.

This paper presents a novel guided genetic algorithm (GGA) for protein folding prediction (PFP) in 2D hydrophobic-hydrophilic (HP) by exploring the protein core formation concept. A proof of the shape for an optimal core is provided and a set of highly probable sub-conformations are defined which help to establish the guidelines to form the core boundary. A series of new operators including diagonal move and tilt move are defined to assist in implementing the guidelines. The underlying reasons for the failure in the folding prediction of relatively long sequences using Unger's genetic algorithm (GA) in 2D HP model are analysed and the new GGA is shown to overcome these limitations. The overall strategy incorporates a swing function that provides a mechanism to enable the GGA to test more potential solutions and also prevent it from developing a schema that may cause it to become trapped in local minima. While the guidelines do not force particular conformations, the result is a number of conformations for particular putative ground energy and superior prediction accuracy, endorsing the improved performance compared with other well established nondeterministic search approaches

Genetic microarray expression data often contains multiple missing values that can significantly affect the performance of statistical and machine learning algorithms. This paper presents an innovative missing value estimation technique, called collateral missing value estimation (CMVE) which has demonstrated superior estimation performance compared with the K-nearest neighbour (KNN) imputation algorithm, the least square impute (LSImpute) and Bayesian principal component analysis (BPCA) techniques. Experimental results confirm that CMVE provides an improvement of 89%, 12% and 10% for the BRCA1, BRCA2 and sporadic ovarian cancer mutations, respectively, compared to the average error rate of KNN, LSImpute and BPCA imputation methods, over a range of randomly selected missing values. The underlying theory behind CMVE also means that it is not restricted to bioinformatics data, but can be successfully applied to any correlated data set.

The Bezier curve is fundamental to many challenging and practical applications, ranging from computer aided geometric design and postscript font representations through to generic object shape descriptors and surface representation. A drawback of the Bezier curve however, is that it only considers global information about the control points, so there is often a large gap between the curve and its control polygon, leading to considerable error in curve representations. To address this issue, this paper presents enhanced Bezier curve (EBC) models which seamlessly incorporate local information. The performance of the models is empirically evaluated upon a number of natural and synthetic objects having arbitrary shape and both qualitative and quantitative results confirm the superiority of both EBC models in comparison with the classical Bezier curve representation, with no increase in the order of computational complexity.

The segmentation performance of any clustering algorithm is very sensitive to the features in an image, which ultimately restricts their generalisation capability. This limitation was the primary motivation in our investigation into using shape information to improve the generality of these algorithms. Fuzzy shape-based clustering techniques already consider ring and elliptical profiles in segmentation, though most real objects are neither ring nor elliptically shaped. This paper addresses this issue by introducing a new shape-based algorithm called fuzzy image segmentation of generic shaped clusters (FISG) that incorporates generic shape information into the framework of the fuzzy c-means (FCM) algorithm. Both qualitative and quantitative analyses confirm the superiority of FISG compared to other shape-based fuzzy clustering methods including, Gustafson-Kessel algorithm, ring-shaped, circular shell, c-ellipsoidal shells and elliptic ring-shaped clusters. The new algorithm has also been shown to be application independent so it can be applied in areas such as video object plane segmentation in MPEG-4 based coding.

Bezier curves (BC) are fundamental to a wide range of applications from computer-aided design through to object shape descriptions and surface mapping. Since BC only consider global information with respect to their control points, this can lead to erroneous shape representations, though integrating local control point information minimises this error. This paper presents a new dynamic Bezier curve (DBC) model which combines both localised and global shape information by making a parametric shift of the BC points in the gap between the curve and its control polygon. The value of the shifting parameter is dynamically determined for a prescribed maximum distortion. DBC retains the kernel properties of the BC without increasing computational complexity order. The model's performance has been empirically evaluated on a number of arbitrary-shaped objects from geometric modelling to shape coding. Both qualitative and quantitative results confirm the improvement achieved compared with the classical BC representation.

Results of any clustering algorithm are highly sensitive to features that limit their generalization and hence provide a strong motivation to integrate shape information into the algorithm. Existing fuzzy shape-based clustering algorithms consider only circular and elliptical shape information and consequently do not segment well, arbitrary shaped objects. To address this issue, this paper introduces a new shape-based algorithm, called fuzzy image segmentation using shape information (FISS) by incorporating general shape information. Both qualitative and quantitative analysis proves the superiority of the new FISS algorithm compared to other well-established shape-based fuzzy clustering algorithms, including Gustafson-Kessel, ring-shaped, circular shell, c-ellipsoidal shells and elliptic ring-shaped clusters.

Proactive self-configuration is indispensable for MANETs like ubiquitous sensor networks (USNs), as component devices of the network are usually exposed to natural or man-made disasters due to the hostile deployment and ad hoc nature of USNs. Network state beacons (NSBs) are exchanged among the key nodes of the network for crucial and effective monitoring of the network for steady state operation. The rate of beacon exchange (F/sub E/) and its contents, define the time and nature of the proactive action. Therefore it is very important to optimize these parameters to tune the functional response of the USN. The paper presents a novel F/sub E/ selection model based on autonomic, load-adaptive optimization of beacon exchange rate for monitoring and proactively reconfiguring the network. The results confirm the improved throughput while maintaining QoS with a little overhead control traffic.

MOTIVATION: Microarray data are used in a range of application areas in biology, although often it contains considerable numbers of missing values. These missing values can significantly affect subsequent statistical analysis and machine learning algorithms so there is a strong motivation to estimate these values as accurately as possible before using these algorithms. While many imputation algorithms have been proposed, more robust techniques need to be developed so that further analysis of biological data can be accurately undertaken. In this paper, an innovative missing value imputation algorithm called collateral missing value estimation (CMVE) is presented which uses multiple covariance-based imputation matrices for the final prediction of missing values. The matrices are computed and optimized using least square regression and linear programming methods. RESULTS: The new CMVE algorithm has been compared with existing estimation techniques including Bayesian principal component analysis imputation (BPCA), least square impute (LSImpute) and K-nearest neighbour (KNN). All these methods were rigorously tested to estimate missing values in three separate non-time series (ovarian cancer based) and one time series (yeast sporulation) dataset. Each method was quantitatively analyzed using the normalized root mean square (NRMS) error measure, covering a wide range of randomly introduced missing value probabilities from 0.01 to 0.2. Experiments were also undertaken on the yeast dataset, which comprised 1.7% actual missing values, to test the hypothesis that CMVE performed better not only for randomly occurring but also for a real distribution of missing values. The results confirmed that CMVE consistently demonstrated superior and robust estimation capability of missing values compared with other methods for both series types of data, for the same order of computational complexity. A concise theoretical framework has also been formulated to validate the improved performance of the CMVE algorithm. AVAILABILITY: The CMVE software is available upon request from the authors.

Very low bit-rate video coding using regularly shaped patterns to represent moving regions in macroblocks has good potential for improved coding efficiency. This paper presents a real-time pattern selection (RTPS) algorithm, which uses a pattern relevance and similarity metric to achieve faster pattern selection from a large codebook. For each applicable macroblock, the relevance metric is applied to create a customized pattern codebook (CPC) from which the best pattern is selected using the similarity metric. The CPC size is adapted to facilitate real-time selection. Results prove the quantitative and perceptual performance of RTPS is superior to both the Fixed-8 algorithm and H.263.

An optimal genetic mutation diagnosis requires proper selection of mutation classifier. This work investigates the performance of different classification, missing value estimation (MVE) and data dimension reduction techniques for the classification of gene expression data for BRCA1, BRCA2 and Sporadic mutations of epithelial ovarian cancer. Bayesian MVE and zero imputation techniques were employed to deal with missing values. Our study showed the better performance of the Bayesian technique. A novel approach is introduced to use generalized regression neural network (GRNN) as genetic mutation classifier which completely outperformed both well established support vector machine and probabilistic neural network.

The protein folding problem is a minimization problem in which the energy function is often regarded as the fitness function. There are several models for protein folding prediction including the hydrophobic-hydrophilic (HP) model. Though this model is an elementary one, it is widely used as a test-bed for faster execution of new algorithms. Fitness computation is one of the major computational parts of the HP model. This paper proposes an efficient search (ES) approach for computing the fitness value requiring only O(n) complexity in contrast to the full search (FS) approach that requires O(n/sup 2/) complexity. The efficiency of the proposed ES approach results due to its utilization of some inherent properties of the HP model. The ES approach represents residues in a Cartesian coordinate framework and then uses relative distance and coordinate polarity to reduce complexity.

Fitness computation after each crossover or mutation operation in genetic algorithm (GA) requires computational time that increases with the increasing length of the chromosome. In this paper, an efficient GA is proposed for protein folding prediction based on the hydrophobic-hydrophilic (HP) model. The partial fitness of the parent computed from one end of sequence till crossover or mutation point is utilized for the computation of the fitness of the child. The calculated value of the partial fitness is stored with the corresponding chromosome. Although the approach requires additional memory for each hydrophobic residue of each chromosome, the computation time is reduced significantly, which is more important than the memory overhead.

This paper presents decision-based fusion models to classify BRCA1, BRCA2 and Sporadic genetic mutations for breast and ovarian cancer. Different ensembles of base classifiers using the stacked generalization technique have been proposed including support vector machines (SVM) with linear, polynomial and radial base function kernels. A generalized regression neural network (GRNN) is then applied to predict the mutation type based on the outputs of base classifiers, and experimental results show that the new proposed fusion methodology for selecting the best and removing weak classifiers outperforms single classification models.

Microarray data often contains multiple missing genetic expression values that degrade the performance of statistical and machine learning algorithms. This paper presents a K ranked diagonal covariance-based missing value estimation algorithm (KRCOV) that has demonstrated significantly superior performance compared to the more commonly used K-nearest neighbour (KNN) imputation algorithm when it is applied to estimate missing values of BRCA1, BRCA2 and Sporadic genetic mutation samples present in ovarian cancer. Experimental results confirm KRCOV outperformed both KNN and zero imputation techniques in terms of their classification accuracies when used toimpute randomly missing values from 1% to 5%.The classifier used for this purpose was the Generalized Regression Neural Network.The paper also provides a hypothesis for why KRCOV performs better than KNN not only for bioinformatics data but also for other data types having strong correlated values.

In the context of very low bit-rate video coding, pattern representations of a moving region (MR) in block-based motion estimation and compensation has become increasingly attractive. Generally, all existing pattern-matching algorithms apply a similarity metric, involving elementary operations, to compute the mismatch between an MR and a particular fixed pattern in order to select the best-matching pattern from a fixed-size codebook of predefined patterns. An efficient similarity metric, together with a new generic computation strategy, is presented by considering only the mismatch areas of MRs. It is theoretically proven that for a specific MR in a macroblock, the new similarity metric selects exactly the same pattern as existing metrics, while the resulting computational coding efficiency is improved by between 21% and 58% compared with the H.263 low bit-rate coding standard.

This paper presents an advanced cell visiting probability estimator for more accurately reserving resources for an estimated time interval to neighbouring cells by exploiting key mobility parameters (speed, direction and distance). The call admission control strategy used allocates bandwidth more efficiently to provide consistent Quality of Service (QoS) guarantees for multimedia traffics. Concomitantly, to ensure continuity of on-going calls with better utilization of resources, bandwidth is borrowed from existing adaptive calls without affecting the minimum QoS guarantee. Simulation results prove that the new scheme offers substantial improvements over the recent existing schemes.

A controlled multicast scheme provides instantaneous service, but limited server bandwidth causes some user requests to be either delayed or rejected when insufficient free bandwidth is available. Two borrowing schemes are proposed for instantaneous video-on-demand (VOD) that reduce the user request blocking rate by borrowing bandwidth from ongoing video streams when there is insufficient free bandwidth for the server to deliver a new video stream. Both these new schemes have proved to be successful in reducing blocking rate and increasing bandwidth utilization at the expense of temporarily degrading the video quality.

This paper proposes a novel velocity support call admission scheme through defining a new directional probability function and a new technique of estimating the reservation time window. The projected velocity-dependent directional probability function, with a negative exponential distribution, is based on the assumption that both the probability of changing direction and the rate of change are higher for slow rather than fast moving mobile units. In this paper, the reservation time window is estimated more accurately through considering the estimated travel distance and average cell dwelling time rather than considering finite number of paths as proposed in many recent studies. The new scheme is compared against the existing predictive mobility support scheme using extensive real-scenario simulations. Results prove the new scheme to be very effective in achieving superior performance in terms of QoS parameters and also reduced number of overhead message transmissions.

Very low bit-rate video coding, using regular shaped patterns to focus on moving regions in macroblocks, has gained significant attention recently. This paper presents a new real-time pattern selection (RTPS) algorithm using a large codebook of thirty two patterns. The algorithm uses a relevance measurement for all the patterns and a moving region, to eliminate a large number of irrelevant patterns prior to the actual best likelihood pattern selection procedure. Both theoretically and empirically it is proven that not only is the computational complexity of the new algorithm comparable to the contemporary algorithm that use a pattern codebook size of only eight patterns but also the new algorithm reduces the bit-rate significantly, while maintaining comparable subjective quality.

Trading-off computational complexity and quality is an important performance constraint for real time application of motion estimation algorithms. To address this issue, a distance dependent thresholding search (DTS) algorithm has been proposed for fast and robust true motion estimation in video coding/indexing applications (Sorwar, G. et al., Proc. ICASSP. 2002; IEEE Asia-Pacific Conf. on Circuits and Systems, 2002; 6th Int. Conf. on Signal Processing, 2002). DTS encompasses both the full search (FS) as well as fast searching modes, with different threshold settings providing various quality-of-service levels. The main drawback of DTS is that the threshold value is defined manually. The DTS algorithm is extended to a fully adaptive distance dependent thresholding search (FADTS), a key feature of which is the automatic adaptation of the threshold using the desired target and the content from the actual video sequence to achieve a guaranteed level of quality or processing complexity. Experimental results confirm the performance of the FADTS algorithm in achieving this objective with minimal additional computational cost.

Most fuzzy rule-based image segmentation techniques to date have been primarily developed for gray level images. In this paper, a new algorithm called fuzzy rule-based colour image segmentation (FRCIS) is proposed by extending the generic fuzzy rule-based image segmentation (GFFUS) algorithm G.C. Karmakar, L.S. Dooley [2002] and integrating a novel algorithm for averaging hue angles. Qualitative and quantitative analysis of the performance of FRCIS is examined and contrasted with the popular fuzzy c-means (FCM) and possibilistic c-means (PCM) algorithms for both the hue-saturation-value (HSV) and RGB colour models. Overall, FRCIS provides considerable improvement for many different image types.

The selection of an optimal regular-shaped pattern set for very low bit-rate video coding, focusing on moving regions has been the objective of much recent research in order to try and improve bit-rate efficiency. Selecting the optimal pattern set however, is an NP hard problem. This paper presents a generic variable pattern selection (GVPS) algorithm, which introduces a pattern selection parameter that is able to control the performance in terms of computational complexity as well as bit-rate and picture quality. While using a sub-optimal variable pattern set, GVPS obtains a coding performance comparable to near-optimal algorithms, such as the k-change neighbourhood solution, while being much less computationally intensive, so that it is able to process all types of video sequences in real-time, with minimal pre-processing overheads.

Very low bit-rate video coding, using regular shaped patterns to focus on moving regions in macroblocks, has gained significant attention. This paper presents a new real-time pattern selection (RTPS) algorithm using a large codebook of thirty two patterns. The algorithm uses a relevance measurement for all the patterns and a moving region, to eliminate a large number of irrelevant patterns prior to the actual best likelihood pattern selection procedure. Both theoretically and empirically it is proven that not only is the computational complexity of the new algorithm comparable to the contemporary algorithm that use a pattern codebook size of only eight patterns but also the new algorithm reduces the bit-rate significantly, while maintaining comparable subjective quality.

The extended fuzzy rules for image segmentation (EFRIS) algorithm initially splits all segmented regions into mutually exclusive 4-connected objects, from which the largest one in each region is designated as its main object. A drawback of this approach is that it is less effective when the main objects are relatively small and some of the minor objects are completely surrounded and connected to the main object of another region. Besides, defining insufficient merging rules, EFRIS also only considers the surrounding main objects in the original order that the regions were segmented, which is undesirable. In this paper, a new general segmentation algorithm called modified extended fuzzy rules for image segmentation (MEFRIS) is presented, which addresses these problems and whose improved segmentation performance is analysed and numerically evaluated. The results are also contrasted with both the original generic fuzzy rule-based image segmentation (GFRIS) and EFRIS algorithms.

A modified full-search (MFS) algorithm is presented for block-based motion estimation applications, which introduces the novel concept of variable distance dependent thresholds. The performance of the MFS algorithm is analyzed and quantitatively compared with both the traditional and exhaustive full-search (FS) technique, and the computationally faster, non-exhaustive three-step-search (TSS) algorithm. Experimental results show that by applying an appropriate threshold function, the MFS algorithm not only matches the speed of the TSS algorithm, but both retains a block distortion error comparable to the global minimum produced by the FS algorithm, and avoids the problem of identifying large numbers of spurious motion vectors in the search process.

The generic fuzzy rule-based image segmentation algorithm (GFRIS) does not produce good results for images containing non-homogeneous regions, as it does not directly consider texture. In this paper a new algorithm called fuzzy rules for image segmentation incorporating texture features (FRIST) is proposed, which includes two additional membership functions to those already defined in GFRIS. FRIST incorporates the fractal dimension and contrast features of a texture by considering image domain specific information. Quantitative evaluation of the performance of FRIST is discussed and contrasted with GFRIS using one of the standard segmentation evaluation methods. Overall, FRIST exhibits considerable improvement in the results obtained compared with the GFRIS approach for many different image types.

We present a classification method based on the discrete cosine transform (DCT) coefficients of texture image. Since the DCT works on gray level images, the color scheme of each image is transformed into gray levels. For classifying the images with DCT, we used two popular soft computing techniques, namely neurocomputing and neuro-fuzzy computing. We used a feedforward neural network trained by backpropagation algorithm and an evolving fuzzy neural network to classify the textures. The soft computing models were trained using 80% of the texture data and remaining was used for testing and validation purposes. A performance comparison was made among the soft computing models for the texture classification problem. We also analyzed the effects of prolonged training of neural networks. It was observed that the proposed neuro-fuzzy model performed better than neural network.

Many fuzzy clustering based techniques do not incorporate the spatial relationships of the pixels, while all fuzzy rule based image segmentation techniques tend to be very much application dependent. In most techniques, the structure of the membership functions are predefined and their parameters are either automatically or manually determined. This paper addresses the aforementioned problems by introducing a general fuzzy rule based image segmentation technique, which is application independent and can also incorporate the spatial relationships of the pixels. It also proposes the automatic defining of the structure of the membership functions. A qualitative comparison is made between the segmentation results using this method and the popular fuzzy c-means (FCM) applied to two types of images: light intensity (LI) and an X-ray of the human vocal tract. The results clearly show that this method exhibits significant improvements over FCM for both types of images.

The generic fuzzy rule-based image segmentation (GFRIS) technique does not produce good results for non-homogeneous regions that possess abrupt changes in pixel intensity, because it fails to consider two important properties of perceptual grouping, namely surroundedness and connectedness. A new technique called extended fuzzy rules for image segmentation (EFRIS) is proposed, which includes a second rule to that defined already in GFRIS, that incorporates both the surroundedness and connectedness properties of a region's pixels. This additional rule is based on a split-and-merge algorithm and refines the output from the GFRIS technique. Two different classes of image, namely light intensity and medical X-rays are empirically used to assess the performance of the new technique. Quantitative evaluation of the performance of EFRIS is discussed and contrasted with GFRIS using one of the standard segmentation evaluation methods. Overall, EFRIS exhibits significantly improved results compared with the GFRIS approach.