Error message

  • Deprecated function: implode(): Passing glue string after array is deprecated. Swap the parameters in drupal_get_feeds() (line 394 of /home/clients/f3577f0844e67df06cec1e182a5384c1/web/includes/
  • Deprecated function: The each() function is deprecated. This message will be suppressed on further calls in menu_set_active_trail() (line 2385 of /home/clients/f3577f0844e67df06cec1e182a5384c1/web/includes/


Distributed Human-based Genetic Algorithm Utilizing A Mobile Ad Hoc Network

Ryosuke Hasebe, Kei Ohnishi and Mario Koeppen

Graduate School of Computer Science and Systems Engineering, Kyushu Institute of Technology,
680-4 Kawazu, Iizuka-shi, Fukuoka 820-8502, Japan


A human-based genetic algorithm (HBGA) is one type of genetic algorithms, in which humans conduct all genetic operators such as selection, crossover, and mutation in a way such that they select others’ solution candidates (selection) and create new candidate solutions influenced by the selected ones (crossover and mutation). HBGA needs a way for people to share their candidate solutions. One way is to manage candidate solutions in a centralized manner as a message board of a web forum, and actually such a HBGA has been implemented. However, how to implement HBGA in a distributed manner has not been well studied so far. This paper presents a method for sharing candidate solutions among humans in HBGA running on a mobile ad hoc network (MANET), which is a distributed system, and shows simulation results to demonstrate the basic usefulness of the proposed method.

Keywords: genetic algorithm, human, ad hoc network, human interest


Scalable Differential Evolution for Many-core and Clusters in Unified Parallel

Pavel Kromer, Jan Platos and Vaclav Snasel

IT4Innovations & Department of Computer Science VSˇB-Technical University of Ostrava,
17. listopadu 12, Ostrava-Poruba, Czech Republic


This study proposes a novel design and implementation of Differential Evolution (DE) using the Partitioned Global Address Space (PGAS) parallel computing model and the Unified Parallel C (UPC) programming language. The mapping of DE concepts to UPC features is presented and a DE useful for both many-core shared memory systems and clusters of computers with distributed memory is implemented and evaluated in the environment of a small real-world high performance computing cluster.

Keywords: Differential Evolution, Parallel Computing, High Performance Computing, Many-core, Cluster, Unified Parallel C.


Constraint Handling in Firefly Algorithm

Aditya Deshpande, Gaurav Phatnani and Anand Kulkarni

Optimization and Agent Technology (OAT) Research Lab, Maharashtra Institute of Technology,
124 Paud Road Pune 411038, India


Most of the contemporary nature-/bio-inspired techniques are unconstrained algorithms. Their performance may get affected when dealing with the constrained problems. There are number of constraint handling techniques developed for these algorithms. This paper intends to compare the performance of the emerging metaheuristic swarm optimization technique of Firefly Algorithm when incorporated with the generalized constrained handling techniques such as penalty function method, feasibility-based rule and the combination of both, i.e. combined approach. Seven well known test problems have been solved. The results obtained using the three constraint handling techniques are compared and discussed with regard to the robustness, computational cost, rate of convergence, etc. The associated strengths, weaknesses and future research directions are also discussed.

Keywords: Metaheuristic, swarm optimization technique, firefly algorithm, penalty function, feasibility-based rule .



Mining Multi-class Industrial Data with Evolutionary Fuzzy Rules

Pavel Kromer, Jan Platos and Vaclav Snasel

IT4Innovations & Department of Computer Science VSˇB-Technical University of Ostrava,
17. listopadu 12, Ostrava-Poruba, Czech Republic


Methods based on fuzzy sets and fuzzy logic have proved to be efficient data classifiers and value estimators. This study presents an application of evolutionary evolved fuzzy rules based on the concept of extended Boolean queries to a multi- class data mining problem. Fuzzy rules are used as symbolic classifiers machine-learned from the data and used to label data samples and predict the value of an output variable. The output variable can be both a label (category) and a continuous value. This study presents an application of evolutionary fuzzy rules to the prediction of multi-class quality attributes in an industrial data set and compares the prediction obtained by fuzzy rules to the prediction achieved by support vector machines.

Keywords: Fuzzy Rules, Genetic Programming, Fuzzy Information Retrieval, Multi-class Data Mining, Industrial Applications.


Combining Expression Trees

Joanna Jedrzejowicz 1 and Piotr Jedrzejowicz2

1 Institute of Informatics, Gdansk University, Wita Stwosza 57, 80-952 Gdansk, Poland
2 Department of Information Systems, Gdynia Maritime University, Morska 83, 81-225 Gdynia, Poland


The paper aims at comparing and evaluating combined classifiers constructed from expression trees induced through applying gene expression programming. Several com- bined classifiers proposed in earlier papers of the authors are briefly described and explained. An extensive computational experiment in which the proposed combined classifiers are com- pared with a number of well-known classification algorithms from WEKA Environment shows that constructing combined classifiers taking advantage of the collective computational techniques brings a very good result in terms of the solution quality.

Keywords: combined classifier, expression tree, gene expression programming.


Dynamic Adaptation of a Vehicle's Cruising Speed with Recurrent
Neural Networks for Enhancing Fuel Economy

Mahmoud Abou-Nasr and Dimitar Filev


This paper presents an architecture for optimally modulating the cruise speed around its set point. Our objective is to minimize the overall fuel consumption over a trip without impacting the overall trip time, by exploiting the vehicle dynamics and the terrain, specifically in this paper, the road grades. The overall trip time is defined as the time to complete the trip while driving at the constant cruise speed, which was set by the driver at the beginning of the trip. We test this architecture with data acquired by an instrumented vehicle driven on city and highway roads in Southeast Michigan. Our testing was very promising and showed that we can achieve up to 11% of overall fuel economy of which 10.8% are from exploiting the road grades.

Keywords: dynamic; adaptation; cruise control; recurrent; neural network; fuel economy; road grades .


A New Graph-based Flooding Matching Method for Ontology Integration

Hai Bang Truong1, Ngoc Thanh Nguyen1, Quoc Uy Nguyen2 and Trong Hai Duong3

1 Computer Science, University of Information Technology, VNU Ho Chi Minh City, Viet Nam ,  2 Institute of Informatics, Wroclaw University of Technology, Poland, 3Faculty of Mathematics and Informatics, Quang Binh University, Viet Nam  


Ontology integration is a well-known problem, a crucial mechanism for semantic interoperability and knowledge reusing, and a backbone of Semantic Web. In this paper, a graph-based method, which combines similarity flooding and concept classification for ontology integration, is proposed. This method consists of three main steps: model ontologies into directed labeled graph, concept classification, and similarity flooding for computing fix-points of pairwise connectivity graph. The main issue presented here is how to shrink spreading scale before we use similarity flooding. Experimental results demonstrate that our method is more effective and obtain better results than original similarity flooding algorithm.

Keywords: Ontologies, Graph matching, OWL, Data graphs, RDF, Integration. 


A Practical Implementation of Self-evolving Cloud-based Control of a Pilot Plant

Bruno Costa1, Igor Skrjanc1, Saso Blazic2 and Plamen Angelov3

1Campus Natal - Zona Norte Federal Institute of Rio Grande do Norte Natal, Brazil
2Faculty of Electrical Engineering University of Ljubljana, Slovenia
3School of Computing and Communications Lancaster University, Lancaster, United Kingdom


This paper presents the implementation of a first order self-evolving cloud-based controller for the liquid level of a two-tank pilot plant. The controller is based on the AnYa type fuzzy rule-based system (FRB), which has a parameter- free antecedent part, and can learn autonomously on-line with each new input data collected and output generated, with no prior knowledge of the system or off-line training. Two types of controllers are considered: a PD-type controller, with simulated and real results; and a MRC-type controller, with simulated results. Regarding the practical implementation, a real continuous process didactic plant was used as a representation of a real industrial environment through the OLE for Process Control (OPC) communication protocol. It has been demonstrated the possibility of building autonomously and in an unsupervised manner a controller capable of developing and adapting itself in a real-time industrial automation application.

Keywords: evolving fuzzy rule-based system, cloud-based controller, industrial processes. 



GMKIT2-FCM: A Genetic-based Improved Multiple Kernel Interval Type-2 Fuzzy C-Means Clustering

Dzung Dinh Nguyen, Long Thanh Ngo and Long The Pham

Le Quy Don Technical University, Hanoi, Vietnam


This paper deals with a Genetic Multiple Kernel Interval Type 2 Fuzzy C-means clustering (GMKIT2-FCM), which automatically find the optimal number of clusters and determine the coefficients of the multiple kernel. The proposed GMKIT2-FCM algorithm provides us a new flexible vehicle to fuse different data information in the classification problems. That is, different information represented by different kernels is combined in the kernel space to produce a new kernel. The proposed algorithm contains two main stages. The first, a heuristic method based on Genetic algorithm (GA) and the average multiple kernel interval type 2 fuzzy c-means clustering (MKIT2-FCM) is adopted to automatically determine the optimal number of clusters and the initial the centroids. Then the results of the first stage are used in combination with GA and MKIT2- FCM to adjust the coefficients of multiple kernel to achieve better results. The experiments are done based on well-known datasets with the statistics show that the algorithm generates good quality of clustering problems.

Keywords: Genetic Algorithm, Type-2 fuzzy sets, type- 2 fuzzy c-means clustering, Multiple kernel-based clustering, Genetic Multiple kernel clustering.


Evolving Markov Chain Models of Driving Conditions Using Onboard Learning

Andrew Hoekstra1, Steve Szwabowski1, Dimitar Filev1, Kevin McDonougha2, and Ilya Kolmanovskya2

1Research and Advanced Engineering, Ford Motor Company Dearborn, Michigan, USA
2Department of Aerospace Engineering University of Michigan, Ann Arbor, Michigan, USA


This paper describes simple and suitable for real-time implementation algorithms for on-board learning of Markov Chain models of driving conditions (e.g., driver wheel torque request, vehicle speed, surrounding traffic speed, road grade, road curvature etc.). The use of Kullback-Liebler (KL) divergence is proposed as a stopping and re-initialization criterion for learning, permitting an evolving set of Markov Chain models to be generated for different route segments. Examples based on learning models of road grade and vehicle speed are reported. Assuming that a set of learned Markov Chain models and of associated control policies is available on- board of the vehicle, the use of KL divergence is also advocated for selecting the control policy that matches the current driving conditions. Potential applications of this approach include optimal energy management in Hybrid Electric Vehicles (HEV) and fuel efficient Adaptive Cruise Control.

Keywords: Markov chain, evolving models, learning, identification, intelligent vehicle control.


Population Learning with Differential Evolution for the Discrete-Continuous Scheduling with Continuous Resource Discretisation

Piotr Jędrzejowicz1 and Aleksander Skakovski2

1Chair of Information Systems, Gdynia Maritime University Gdynia, Poland
2Department of Navigation, Gdynia Maritime University Gdynia, Poland


In the paper, we consider a population learning algorithm denoted (PLA3), with the differential evolution method for solving the discrete-continuous scheduling problem (DCSP) with continuous resource discretisation - ΘZ. The considered problem originates from DCSP, in which nonpreemtable tasks should be scheduled on parallel identical machines under constraint on discrete resource and requiring, additionally, a renewable continuous resource to minimize the schedule length. The continuous resource in DCSP is divisible continuously and is allocated to tasks from a given interval in amounts unknown in advance. Task processing rate depends on the allocated amount of the continuous resource. To eliminate time consuming optimal continuous resource allocation, an NP-hard problem ΘZ with continuous resource discretisation is introduced and sub- optimally solved by PLA3. Experimental results show that PLA3 was able to improve best-known solutions and excels its predecessor PLA2 in solving the considered problem.

Keywords: Population Learning Algorithm, Differential Evolution Method, Discrete-Continuous Scheduling, Discretisation .


Intelligent Semantic Question Answering System

Erfan Najmi, Khayyam Hashmi, Fayez Khazalah and Zaki Malik

Department of Computer Science, Wayne State University, Detroit, Michigan, USA


The volume of information available on the World Wide Web and the rate of its growth requires new techniques to handle and organize this data. Ontologies are becoming the pivotal methodology to represent domain-specific conceptual knowledge and hence help in providing solutions for Question Answering (QA) systems. This paper introduces an approach for enhancing the capabilities of QA systems using semantic technologies. We implemented an approach to convert the natural language user queries to Resource Description Framework (RDF) triples and find relevant answers. The experiment results show that the proposed technique works very well for single word answers. We believe that with some modifications this approach can be expanded to a wider scale.

Keywords: answering system, semantic.



Consensus-based Cluster Merging for the Prototype Selection

Ireneusz Czarnowski and Piotr Jędrzejowicz

Department of Information Systems, Gdynia Maritime University, Morska 83, 81-225 Gdynia, Poland


The aim of the paper is to propose and evaluate a hybrid approach to generate a representative training dataset of the required size. Prototype selection is understood as a selection of the representative prototypes from the original training dataset. The basic assumptions underlying the proposed method is that the prototype selection is carried out after the training dataset has been grouped into clusters, and that prototypes are selected from each of thus obtained clusters. Under these assumptions the number of clusters produced has a direct influence on the size of the reduced dataset. When the number of clusters exceeds the required final size of the training dataset, clusters need to be merged. Clusters merging may not be an easy task in case clusters have a heterogeneous structure. The paper considers the problem of cluster merging and proposes to eliminate the problem of the cluster heterogeneity through reaching a consensus-based solution.

Keywords: Consensus-based, Cluster Merging.


The Effectiveness of using Geometrical Features for Facial Expression Recognition

Anwar Saeed, Ayoub Al-Hamadi and Robert Niese

Institute for Electronics, Signal Processing and Communications (IESK),
Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, P.O. Box 4210 Germany


Facial expressions play an important role in diverse disciplines ranging from entertainment (video games) to medical applications and affective computing. For tackling the problem of expression recognition, various approaches were proposed over the last two decades. These approaches are primarily divided into two types: geometry and appearance based. In this paper, we address the geometry based approaches to recognize the six basic facial expressions (happiness, surprise, anger, fear, disgust, and sadness). We provide answers to three major questions regarding the geometrical features: 1. What is the minimum number of facial points that could provide a satisfactory recognition rate? 2. How this rate is affected by prior knowledge of person- specific neutral expression? 3. How accurate should a facial point detector be to achieve an acceptable recognition rate? To assess the reliability of our approach, we evaluated it on two public databases. The results show that a good recognition rate could be achieved by using just eight facial points. Moreover, the lack of prior knowledge of person-specific neutral state causes more than 7% drop in the recognition rate. Finally, the recognition rate is adversely affected by the facial point localization error.

Keywords: Facial Expressions, Fiducial Facial Points, Sup- port Vector Machine.


Multi-View Video Based Tracking and Audio-Visual Identification of Persons in a Human-Computer-Interaction Scenario

Sascha Meudt, Michael Glodek, Martin Schels and Friedhelm Schwenker

Institute of Neural Information Processing, University of Ulm, Germany


User identification and tracking are definitely the basic tasks in any human computer interaction (HCI) scenario. For these tasks we propose a multi-view approach utilizing multi- camera systems and audio processing systems. Face detectors and face recognizers are based on orientation histogram and eigenface techniques, and Mel Frequency Cepstral Coefficients (MFCC) are applied for speaker identification. In order to achieve a robust user identification and localization spatio-temporal classifier fusion methods have been integrated into the overall classifier system, support vector machines (SVM) and k nearest neighbor (kNN) models are used as base classifiers. A general office environment with up to six persons was the test bed for data collection and numerical evaluation.

Keywords: Video Tracking, Person Identification, Human Position Estimation, Human Computer Interaction, Speaker Identification.


Ethical Considerations for Engineers Working in Cybernetic Implants

Kate Fox

School of Physics, University of Melbourne, Parkville, Victoria, Australia


Research, engineers and ethics are not commonly synonymous. Cybernetic research and particularly neural implant systems exist on the spectra of ethics, especially considering the potential outcome of the research. Neural implants have shown considerable advantages for the repair and replacement of damaged or disabled biological systems and accordingly are rapidly being introduced into mainstream medicine. Concerns however exist over prerequisite preclinical testing and potential misuse of the technology.

Keywords: engineering; neural implants; ethics; medicine.


Advancement and Systematic Validation of an Automated Pain Recognition System

Steffen Walter1, Sascha Gruss1, Jun-Wen Tan1, Hagen Ehleiter1, Harald Traue1,
Philipp Werner2,  Ayoub Al-Hamadi2, Stephen Crawcour3,
Adriano O. Andrade4 and Gustavo Moreira Da Silva4

1Department Psychosomatic Medicine and Psychotherapy, University of Ulm, Ulm, Germany
2Clinical Psychology and Psychotherapy, Technical University of Dresden, Dresden, Germany
3Institute for Electronic, Signal Processing and Communication, University of Magdeburg Magdeburg, Germany
4Biomedical Engineering Laboratory (BioLab), Federal University of Uberlandia, Uberlandia, Brazil


The objective measurement of subjective, multi- dimensionally experienced pain is still a problem that has yet to be adequately solved. Though verbal methods (i.e., pain scales, questionnaires) and visual analogue scales are commonly used for measuring clinical pain, they tend to lack in reliability or validity when applied to mentally impaired individuals. Expression of pain and/or its biopotential parameters could represent a solution. While such coding systems already exist, they are either very costly and time-consuming, or have been insufficiently evaluated with regards to the theory of mental tests. Building on the experiences made to date, we collected a database using visual and biopotential signals to advance an automated pain recognition system, to determine its theoretical testing quality, and to optimize its performance. For this purpose, participants were subjected to painful heat stimuli under controlled conditions.

Keywords: pain; quantification; heat; biopotential; facial expression; pain computing .


Using Speaker Group Dependent Modelling to Improve Fusion of Fragmentary Classifier Decisions

Ingo Siegert1, Michael Glodek2, Axel Panning3, Gerald Krell3,
Friedhelm Schwenker2, Ayoub Al-Hamadi3 and Andreas Wendemuth1

1Cognitive Systems Group, Otto von Guericke University Magdeburg, Germany;
2Institute of Neural Information Processing, Ulm University, Germany;
3Technical Computer Science Group, Otto von Guericke University Magdeburg, Germany


Current speech-controlled human computer inter- action is purely based on spoken information. For a successful interaction, additional information such as the individual skills, preferences and actual affective state of the user are often mandatory. The most challenging of these additional inputs is the affective state, since affective cues are in general expressed very sparsely. The problem can be addressed in two ways. On the one hand, the recognition can be enhanced by making use of already available individual information. On the other hand, the recognition is aggravated by the fact that research is often limited to a single modality, which in real-life applications is critical since recognition may fail in case sensors do not perceive a signal. We address the problem by enhancing the acoustic recognition of the affective state by partitioning the user into groups. The assignment of a user to a group is performed at the beginning of the interaction, such that subsequently a specialized classifier model is utilized. Furthermore, we make use of several modalities, acoustics, facial expressions, and gesture information. The combination of decisions not affected by sensor failures from these multiple modalities is achieved by a Markov Fusion Network. The proposed approach is studied empirically using the LAST MINUTE corpus. We could show that compared to previous studies a significant improvement of the recognition rate can be obtained.

Keywords: Multimodal Pattern Recognition, Affect Recog- nition, Companion Systems, Human Computer Interaction .


Is Natural Language Ever Really Vague: A Computational Semantic View

Victor Raskin, Julia Taylor and Lauren Stuart

Purdue University, West Lafayette, Indiana, USA


The paper starts out with an observation that, in the domain of fuzzy logic, fuzzy sets, computing with words, etc., the charges from the outside that fuzziness equals probability are routinely and calmly rebuffed, but confusing fuzziness with vagueness has not been ultimately dealt with even inside the community. We leave completely aside the category of vagueness that is an artifact of approaches, both in logic and philosophy as well as trends in linguistics, such as formal semantics, that attempt to apply predicate logic of various flavors and complexity to a limited selection of language phenomena, such as quantifiers and scalars that lend themselves to such a treatment. Instead, using a computational semantic approach based on a language- independent ontology and language-specific lexicons, where each entry is anchored in and defined with the help of ontological properties and concepts, the paper claims that, unlike fuzziness, vagueness is not an inherent feature of certain words, phrases, or sentences. In fact, it is suggested that vagueness does not really exist for a human hearer and thus is just a temporary function of discourse, in which the speaker's grain size level is coarser than that of the hearer. Since hearers handle it routinely by asking for more details, the paper outlines the computational procedure emulating this ability.

Keywords: fuzziness; computational semantics; ontological semantic technology; ontology; lexicon .


Simulated Annealing Approach to Fuzzy Modeling of Servo Systems

Radu-Emil Precup1, Mircea-Bogdan Radac1, Emil Petriu1, Claudia-Adina Dragos1 and Stefan Preitl2

1Department of Automation and Applied Informatics, "Politehnica" University of Timisoara Timisoara, Romania
2School of Electrical Engineering and Computer Science University of Ottawa, Ottawa, Canada


This paper proposes an approach to the fuzzy modeling of servo systems using Simulated Annealing (SA) algorithms. A set of local state-space models is obtained from the first principle models of the process. The initial Takagi-Sugeno- Kang (TSK) fuzzy models are obtained by the modal equivalence principle, where the local state-space models are placed in the rule consequents. Optimization problems are defined aiming the minimization of objective functions expressed as integrals of squared modeling errors. The variables of the objective functions are the limits of the supports of the input membership functions and the kernels of these membership functions are kept constant. SA algorithms are implemented to solve the optimization problems which yield optimal TSK fuzzy models. A set of real- time experimental results for a laboratory nonlinear servo system validates the new optimal TSK fuzzy models.

Keywords: fuzzy models; modal equivalence principle; servo systems; Simulated Annealing; optimization .


A Hybrid Computational Chemotaxis in Bacterial Foraging Optimization Algorithm for Global Numerical Optimization

Yosra Jarraya1, Souhir Bouaziz1, Adel M. Alimi1 and Ajith Abraham2,3

1REsearch Group on Intelligent Machines (REGIM), University of Sfax, National School of Engineers (ENIS), BP 1173, Sfax 3038, Tunisia 2Machine Intelligence Research Labs, WA, USA 3IT4Innovations, VSB-Technical University of Ostrava, Czech Republic


This paper first proposes a simple scheme for adapting the chemotactic step size of the Bacterial Foraging Optimization Algorithm (BFOA), and then this new adaptation and two very popular optimization techniques called Partic le Swarm Optimization (PSO) and Differential Evolution (DE) are coupled in a new hybrid approach named Adaptive Chemotactic Bacterial Swarm Foraging Optimization with Differential Evolution Strategy (ACBSFO_DES). This novel technique has been shown to overcome the problems of premature convergence and slow of both the classical BFOA and the other BFOA hybrid variants over several benchmark problems.

Keywords: Adaptive Bacterial Foraging Optimization Algorithm, Particle Swarm Optimization, Differential Evolution, Hybrid Computational Chemotaxis.


Improvements in Feature Vector Selection and Parameter Optimisation for Continuous Gesture Recognition

Ashley Gritzman, Tomislav Batev and Adam Pantanowitz

Biomedical Engineering Research Group, School of Electrical and Information
Engineering University of the Witwatersrand, Johannesburg, South Africa


Gesture recognition has attracted significant interest due to diverse potential applications, including: hand writing recognition, robot control and human-computer interfaces. This paper identifies and addresses three shortcomings in current approaches to feature vector selection and parameter optimisation for continuous gesture recognition. First, in selecting the final feature vector, researchers typically analyse only a small subset of possible feature combinations; however, the limited subset is likely to omit the optimum feature vector. Second, selection of the final feature vector is based on performance in isolated recognition; however, the final feature vector may not perform adequately in continuous recognition. No protocol currently exists to evaluate and select the final feature vector in continuous recognition mode, thus a novel scoring system is developed. Finally, optimisation of the number of states in the Hidden Markov Models (HMMs) and the number of clusters (k- means clustering) is performed independently, ignoring any possible interdependency. To investigate and address these shortcomings, a gesture recognition system geared towards sign language interpretation is designed. The system is tested on a 9- word gesture vocabulary, and subsequent analysis confirms the above conjectures: first, the optimum feature vector cannot be intuitively predicted and must be determined through rigorous analysis; second, selecting the final feature vector in continuous mode improved the accuracy score by 5.85 % and the perfect sentence recognition by 47.2 %; finally, optimising the number of states and number of clusters simultaneously improved the accuracy score by 3.0 % and the perfect sentence recognition by 11.1 %.

Keywords: gesture recognition; feature vector selection; Hidden Markov Models; parameter optimisation; Kinect Sensor .


Extended Immune Programming and Opposite-based PSO for Evolving Flexible Beta Basis Function Neural Tree

Souhir Bouaziz1, Adel M. Alimi1 and Ajith Abraham2,3

1REsearch Group on Intelligent Machines (REGIM), University of Sfax, National School of Engineers (ENIS), BP 1173, Sfax 3038, Tunisia
2 Machine Intelligence Research Labs, WA, USA
3 IT4Innovations, VSB-Technical University of Ostrava, Czech Republic


In this paper, a new hybrid learning algorithm based on the global optimization techniques, is introduced to evolve the Flexible Beta Basis Function Neural Tree (FBBFNT). The structure is developed using the Extended Immune Programming (EIP) and the Beta parameters and connected weights are optimized using the Opposite-based Particle Swarm Optimization (OPSO) algorithm. The performance of the proposed method is evaluated for time series prediction area and is compared with those of associated methods.

Keywords: Extended Immune Programming; Opposite-based Particle Swarm Optimization; Flexible Beta Basis Function Neural Tree; Time series prediction.


Making the Most of Context-awareness in Brain-computer Interfaces

Sareh Saeedi, Tom Carlson, Ricardo Chavarriaga and José Del R. Millán

Chair in Non-Invasive Brain-Machine Interface (CNBI), Center for Neuroprosthetics,
School of Engineering, EPFL Lausanne, Switzerland


In order for brain-computer interfaces (BCIs) to be used reliably for extended periods of time, they must be able to adapt to the users evolving needs. This adaptation should not only be a function of the environmental (external) context, but should also consider the internal context, such as cognitive states and brain signal reliability. In this work, we propose three different shared control frameworks that have been used for BCI applications: contextual fusion, contextual gating, and contextual regulation. We review recently published results in the light of these three context-awareness frameworks. Then, we discuss important issues to consider when designing a shared controller for BCI.

Keywords: brain-computer interface (BCI), shared control, adaptive assistance .


Dependency Modeling in Automated Web Services Quality Component Negotiations

Khayyam Hashmi, Amal Alhosban, Erfan Najmi and Zaki Malik

Department of Computer Science, Wayne State University, Detroit, Michigan, USA


Web services are a popular choice for component oriented systems that support dynamic compositions. Automated negotiation among Web services provides an effective way for the services to bargain for their optimal customizations and allows the discovery of overlooked potential solutions. Unique and dynamic Quality of Service (QoS) requirements of service consumers pose challenges for effective approaches of service compositions with multiple QoS parameters. In this paper, we present a negotiation Web service that would be used by both the consumer and provider Web services for conducting negotiations for dependent QoS parameters. We use a genetic algorithm(GA) based approach for finding acceptable solutions in multi-party and multi-objective scenarios.Experimental results indicate the applicability and improved performance or our approach in facilitating the negotiations involved in a Web service composition process.

Keywords: web services.


A Rule-Plus-Exemplar Classification System for Adapting to Concept Growth

Wing Yee Sit and Kezhi Mao

School of Electrical and Electronic Engineering Nanyang Technological University Singapore


This paper proposes a rule-plus-exemplar classification system to deal with the concept growth problem. Unlike concept drift, the concept is expanding with time rather than becoming obsolete. The proposed system is able to grow and evolve to incrementally learn the concept. It also adapts to the change to provide reliable classification even when the sample is unfamiliar with respect to the available training data. A series of experimental results with comparable methods show that the system can perform better under concept growth circumstances.

Keywords: concept growth; incremental learning; pattern classification; underrepresented concept .


Anonymization: A Tool for Anonymization of the Internet Traffic

Tanjila Farah and Ljiljana Trajkovic

Simon Fraser University Vancouver, British Columbia, Canada


Collecting network traffic traces from deployed net- works is one of the basic steps in understanding communication networks. Traffic traces are used for network management, traffic engineering, packet classification, and analyzing user behavior to ensure adequate quality of service. Monitored traffic traces should be anonymized for privacy and security reasons. The goal of anonymization is to preserve trace properties while enforcing privacy policies. Various tools and techniques have been implemented for trace anonymization. In this paper, we pro- pose and implement an anonymization tool that executes multi- level anonymization and displays analysis results. We describe architecture and features of the tool and discuss analysis of un- anonymized and anonymized datasets.

Keywords: Network traffic, anonymization, traffic analysis.


Clinical Schedule Management based on Granularity-based Mining

Shusaku Tsumoto1, Haruko Iwata1 and Shoji Hirano2

1Division of Nursing, Shimane University Hospital, 89-1 Enya-cho Izumo, Shimane 693-8501, Japan
2Department of Medical Informatics, School of Medicine, Shimane University, 89-1 Enya-cho Izumo, Shimane 693-8501 Japan


Schedule management of hospitalization is impor- tant to maintain or improve the quality of medical care and application of a clinical pathway is one of the important solutions for the management. Although several kinds of deductive methods for construction for a clinical pathway have been proposed, the customization is one of the important problems. This research proposed an inductive approach to support the customization of existing clinical pathways by using data on nursing actions stored in a hospital information system. Since hospital data include temporal trends of clinical symptoms and medical services, we can discover not only knowledge about temporal evolution of disease, but also one about medical practice from hospital information system. This paper proposes temporal data mining process and applied the method to capture temporal knowledge about nursing practice. The results show that the reuse of stored data will give a powerful tool for management of nursing schedule and lead to improvement of hospital services.

Keywords: temporal data mining; clustering; multidimensional scaling; hospital information system; visualization.


Cybernetics: Where Shall We Go?

Qiangfu Zhao1, John Brine1 and Dimitar P. Filev2

1Dept. of Computer and Information Systems The University of Aizu Aizuwakamatsu, 965-8580, Japan
2Research & Innovation Center Ford Motor Company, 2101 Village Rd., Dearborn, MI 48121


Cybernetics, as defined by Plato and later by Ampère, is the science of governance. In the 1940s, Wiener used cybernetics as an umbrella term to refer to control and communication in both the animal and the machine. In the following decades, the term has been defined in various ways by different researchers, and because of this, cybernetics has been perceived rather negatively as a “nomad science”. Consequently, few people understand the true meaning of cybernetics. For the appropriate development of our field of research, we think it is necessary to re-consider the meaning and the scope of cybernetics, so that we can have a relatively clear mission in our research. In this paper, we try to provide a kind of governance message that might also be very weak, but nevertheless may be helpful for the cybernetics community to become cybernetic itself.

Keywords: ybernetics; governance; self-governance; artificial intelligence; awareness; search; evolution.


Linger Thermo Theory, Part I: The Dynamics Dual of the Stationary Entropy-Ectropy Based Latency Information Theory

Erlan H. Feria

Department of Engineering Science and Physics, City University of New York/CSI, USA


A statistical-physics and information-systems based linger thermo theory is advanced that is the dynamics dual of the stationary entropy-ectropy based latency information theory. It addresses operating issues of information sources, retainers, processors and movers that are contained in a closed-system, or universe, and whose solutions are enabled by a novel unifying duality language. Linger thermo theory combines a newly enhanced thermodynamics, which addresses both information- source’s order and information-retainer’s retention issues, with its recently discovered time dual, called lingerdynamics that is concerned with information-processor’s connection and information-mover’s mobility issues. The theory is a realistic predictor of wide ranging phenomena. Among these one finds: 1) that a closed-system, or universe, continuously expands; 2) that the theoretical life expectancy of an adult living system can be mass independent, an unexpected and surprising 2010 linger thermo prediction strongly supported by a lifespan study started in the 1980s of rhesus monkeys by the United States National Institute of Aging (NIA) whose results were published in a 2012 Nature article: the NIA investigators were shocked with these results since they actually aimed to show that the life expectancy of higher mass (obese) rhesus monkeys was significantly less than that of lower mass (non-obese) ones; and 3) an equation that predicts our perceived faster moving of time as we age.

Keywords: thermodynamics, lingerdynamics, entropy, ectropy, information, latency, statistical physics, lifespan, cosmology, time compression, biology, biochemistry .


Modeling the Evolution of Post Disaster Social Awareness from Social Web Sites

Basabi Chakraborty1 and Soumya Banerjee2

1Faculty of Software and Information Science, Iwate Prefectural University, Takizawa,Iwate, Japan
2Department of Computer Science, Birla Institute of Technology, Mesra, India


With the rapid growth of information and communi- cation technologies, varieties of social media and web based social networks are rapidly emerging. The vast amount of information quickly spreading through social networks provides tremendous challenges and opportunities for researchers trying to gain insight into patterns of human interaction and collective behavior. In this work a computational framework for modeling the evolution of social awareness after an event of manmade or natural disaster by analysis of social media data is proposed. Here we examined the evolution of the social dynamics, sentiment, opinion and views of the people after a major event from web blog messages and tried to build a model of the behavioral pattern by swarm intelligence, a computational intelligence technique. A simple simulation experiment has been done with facebook and blog messages after Mumbai terrorist attack in India in 2011.

Keywords: Social awareness, social web sites, sentiment flow, swarm intelligence .


Linger Thermo Theory, Part II: A Weight Unbiased Methodology for Setting Life Insurance Premiums

Erlan H. Feria

Department of Engineering Science and Physics, City University of New York/CSI, USA


A nascent linger thermo theory is found to lead to a weight unbiased methodology for setting life insurance premiums. The approach is based on a theoretical adult lifespan τ calculated according to:⎛ M ⎞2 τ = Δ τ ⎜⎝ Δ M ⎟⎠where M is the adult’s mass, ΔM is the daily food mass (e.g., 0.4 kg for a daily 2,000 kcal diet of a M=70 kg adult) and Δτ is the duration of one day. Since τ is proportional to the ratio of an individual’s mass to the consumed food per day squared, i.e., (M/ΔM)2, it predicts that the theoretical adult life expectancy of an individual can be weight independent as long as the ratio M/ΔM remains constant as he gains or loses fat. Most importantly, this 2010 theoretical prediction is supported by United States National Institute of Aging (NIA) rhesus monkey study results, first reported in a 2012 Nature journal article, which surprised and shocked the researchers when they discovered that higher weight (obese) monkeys had a similar life expectancy as lower weight ones. It is thus expected that the proposed premium acquisition method should improve on traditional calculations and actuarial tables that often presume that obese individuals have lower life expectancies.

Keywords: thermodynamics, lingerdynamics, entropy, ectropy, information, latency, statistical physics, lifespan, life insurance, premium, actuarial tables, weight .


An Intelligent State Machine towards Task-Oriented Search Support

Neil Y. Yen1, Qiangfu Zhao1, Yong Liu1 and Joseph C. Tsai2

1 School of Computer Science and Engineering The University of Aizu Aizu-Wakamatsu, Japan
2Foundation of Computer Science Laboratory, The University of Aizu Aizu-Wakamatsu, Japan


Researchers tend to agree that an increasing quantity of data has caused the complexity and difficulty for information discovery, management and reuse. An essential factor relates to the increasing channels for information sharing. Finding information, especially those meaningful or useful one, that meets ultimate task of user becomes harder then it is used to be. In this research, issues concerning the use of user-generated contents for individual search support are investigated. In order to make efficient use of user-generated contents, an intelligent state machine, as a hybridization of graph model and petri-net model (i.e., Document Sensitive Petri-Net), is proposed. It is utilized to clarify the vague usage scenario between user-generated contents, such as discussions, posts, etc., and to identify correlations and experiences within them. As a practical contribution, an interactive search algorithm that generates potential solutions for individual is implemented. The feasibility of this research is demonstrated by a series of experiments and empirical studies with around 320,000 user-generated contents collected from the Internet and 180 users.

Keywords: Intelligent State Machine, Document Sensitive Petri- Net, Human-Centered, Decision Support, User-Generated Contents .


Semantic Medical System for Health Condition Awareness

Jie Ji1, Peter Scholten2, Liang Chen3 and Qiangfu Zhao4

1Computer Science Department, Jining University, Qufu, China
2Scholten Consultancy, Netherland
3Association for Friendship with Foreign Countries Jining government, Jining, China
4System Intelligence Lab, The University of Aizu, Aizu-wakamatsu, Japan


Along with the global aging problem, coming years special attention has to be given to ICT for Health and ageing well with his focus on Prevention, Health promotion and integrated Health Care. We have developed a semantic medical information system. The system could collect medical information from the Internet, convert knowledge to ontology and answer user ques- tions. In this paper, we proposed two components for healthcare improving and disease prevention. With an awareness component the system could understand user’s condition, and a feedback components motivates users to adapt their behavior from sickness to health.

Keywords: Awareness computing, healthcare, medicinal information, ontology, semantic web, pattern recognition .


New Haze Removal Scheme and Novel Measure of Enhancement

Sos Agaian and Mehdi Roopaei

Department of Electrical Engineering, University of Texas at San Antonio, USA


Present article addresses novel method to get rid of haze effect in color and gray image enhancement. Proposed scheme works for different hazing cases such as smoky and foggy images. The basic steps of the presented algorithm are: at first an optimized histogram mapping function is applied on original images. In the second step, every channel is added with a combination of original and the related filtered channel. A coordinate transformation from RGB color model to HSV and applying CLAHE enhancement on the v-channel would be assigned at the end of procedure. The key advantages of the proposed algorithm are: it is very simple and based on contrast and color enhancement. Besides, in removal of haze from color and gray images, there is no need to consider any mathematical model for the airlight. In this article, a novel measurement for image enhancement is also introduced. The proposed measure attempts to use all data instead of using just max and min in local blocks obtained by image splitting. Comparisons with well- known NASA Retinex algorithm and recently published methods have shown that our achievements are more effective in removal haze image enhancement.

Keywords: Image Enhancement-Haze Effect –Measure of Enhancement.


Empathetic Healthcare Services based on U-pillbox System

Xin Zhao1, Toshihiro Tamura1, Runhe Huang2 and Jianhua Ma2

1Graduate School of Computer and Information Sciences, Hosei University, Tokyo, Japan
2Faculty of Computer and Information Sciences, Hosei University, Tokyo, Japan


On the basis of advanced network and sensor technology, u-healthcare (ubiquitous healthcare) systems, which reshape the traditional healthcare systems and enhance e-healthcare (electronic healthcare) systems with awareness of a patient’s situation, are the promising healthcare service provision systems. The empathetic personalized u-pillbox system is a heuristic example of u- healthcare systems that emphasizes humanistic healthcare service provision to the elderly. This paper briefly describes a u-pillbox system, which takes into consideration the characteristics of users, demonstrates a home server based system design, and shows the simulation of how the u-pillbox provides humanistic medication dispensing services. Finally it is addressed that how the system can be extended for potential widely applications such as further healthcare services to elderly users, their family members or other concerned parties.

Keywords: empathetic healthcare; u-pillbox; simulation system; personalized service; u-healthcare .


No Reference Color Image Quality Measures

Chen Gao1, Karen Panetta1 and Sos Agaian2

1Electrical and Computer Engineering Department Tufts University, Medford, MA, USA
2Electrical and Computer Engineering Department , University of Texas at San Antonio, San Antonio, TX, USA


Color image quality assessment is essential in evaluating the performance of color image enhancement and retrieval algorithms. Much effort has been made in recent years to develop objective image quality metrics that correlate with perceived quality measurements. Unfortunately, only limited success has been achieved [1]. In this paper we present: a) a new contrast based grayscale image quality measure: Root Mean Enhancement (RME); b) a color RME contrast measure CRME which explores the three dimensional contrast relationships of the RGB color channels; c) a color measure Color Qiality Enhancement (CQE) which is based on the linear combination of colorfulness, sharpness and contrast. Computer simulations show that the new measures may help to evaluate color image quality and choose the optimal operating parameters in color image processing systems. We demonstrate the effectiveness of the presented measures by using the TID2008 database. We also compare the presented measures with subjective evaluation Mean Opinion Score (MOS). Experimental results show good correlations between the presented measures and MOS.

Keywords: color measure, Root Mean Enhancement (RME), Color/Cube Root Mean Enhancement (CRME), Color Quality Enhancement (CQE).


Human Awareness Viewed from Natural Language Concept Formation

Masao Yokota

Faculty of Information Technology Fukuoka Institute of Technology Fukuoka, Japan


Mental Image Directed Semantic Theory (MIDST) has already shown that each physical event concept (e.g., “carry”, “separate”) in natural language is characterized by a so-called “event pattern”, abstract pattern formed by the constituents of its referents. Therefore, people are assumed significantly aware of the event pattern involved when they cognize or recognize a physical event discerned with the others. Such event patterns are modeled as so-called “loci in attribute spaces” in MIDST. This is also the case for mental event concepts (e.g., “love”, “sympathize”). For example, a 5-dimensional attribute space can be provided for human emotion. This paper describes an approach toward a human mentality, so called Kansei, in order to provide robots with a function to measure peoples’ emotions toward external things, focusing on human awareness in concept formation of affective words related to facial expressions of Buddhism statues.

Keywords: natural language; concept formation; emotion


Sparse Representation Using Contextual Information for Hyperspectral Image Classification

Haoliang Yuan, Yang Lu, Lina Yang, Huiwu Luo and Yuanyan Tang

Department of Computer and Information Science, University of Macau, Macau  


This paper analyzes the classification of hyperspec- tral images with the sparse representation algorithm in the presence of a minimal reconstruction error. Incorporating the contextual information into the sparse recovery process can improve the classification performance. However, previous sparse algorithms using contextual information only assume that all neighbors around a test sample make equal contributions to the classification. One disadvantage is that these neighbors located in the edge may belong to the different classes, because they are extracted by a fixed square window. Assuming equal contributions may ease the discrimination of the obtained sparse represen- tations. In this paper, we propose a least square based sparse representation algorithm, which uses the weight vector obtained by the least square method from the neighbors to help improve the sparse representations. Through projecting the weight vector into the corresponding sparse representations, the obtained sparse representations can build a relationship between the neighbors through different weights. Comparative experimental results are shown to demonstrate the validity of our proposed algorithm.

Keywords: Hyperspectral image, classification, sparse rep- resentation, least square.


Awareness of Social Influence on Linked Social Service

Wuhui Chen, Incheon Paik, B.T.G.S Kumara, Takazumi Tanaka and Junbo Wang

School of Computer Science and Engineering, University of Aizu, Aizu-Wakamatsu, Fukushima, Japan


Linked social service considers not only the functionality and QoS of service but also the service’s sociability, so that it knows not only about themselves, but also the peers that they would like to work with in case of composition or they would have to compete against in case of service selection. Global social service network was constructed by connecting linked social services, to describe service societies’ features such as social relations and social states, and provide a basis for inferring, planning, and coordinating social activities. Therefore, awareness of the social relationship between linked social services can help many mining applications such as representative node identification and service recommendation. In this paper, we propose a flexible model for effective awareness of social influence to provide a quantitative measure of the influential strength. First, we formally formulate the problem of awareness of social influence in general domains; next, we observe some fundamental social factors which impact the social influence strength between Linked social services in global social service network; and then, a flexible model is proposed for awareness of social influence on Linked social service to provide a quantitative measure of social influence strength. Finally, an application examples, such as representative service identification is provided.

Keywords: Linked social service; global social service network; social factor; social influence; representative node identification.


Spectral-Spatial Linear Discriminant Analysis for Hyperspectral Image Classification

Haoliang Yuan, Yang Lu, Lina Yang, Huiwu Luo and Yuanyan Tang

Department of Computer and Information Science, University of Macau, Macau


We propose a spectral-spatial linear discriminant analysis method (LDA) for dimensionality reduction in hyper- spectral image. The proposed method uses a local scatter of the small neighborhood as a regularizer to incorporate into the objective function of the LDA. The intrinsic idea is to design an optimal linear transformation that makes these samples among the neighborhood approximate the local mean in the low-dimensional feature space while simultaneously preserving the original property of LDA. Experimental results based on both adequate training samples and inadequate training sam- ples demonstrate that the proposed method outperforms several traditional dimensionality reduction methods.

Keywords: Hyperspectral image, classification, linear dis- criminant analysis, spectral-spatial


Perception Delay and its Estimation Analyzing EEG Signal

Goutam Chakraborty1, Daigo Kikuchi2, Jun Sawamoto1, Hikaru Yokoha2

1Dept. of Software & Information Science Iwate Prefectural University, Iwate, Japan
2Grad. School of Software & Information Science Iwate Prefectural University, Iwate, Japan


Human brain receives a variety of raw data through sense organs, and process them to generate meaningful infor- mation like an image or an alarm etc. These environmental information is passed to and integrated in cerebral cortex, where it is evaluated with respect to knowledge acquired from previous experiences. Finally, if the situation warrants an effort for a reward or survival, a strong signal is generated. This signal activates primary motor cortex for movement and control of limbs, thus playing a vital role in subsequent action. We call this Perception, mechanism to be Aware. A bottom-up processing of the raw sensor modalities to form an abstraction of the environment, combined with a top-down processing of the memory and knowledge acquired through previous experiences, are involved in the complex process of Perception.

Keywords: EEG, Wavelet transform, Clustering, Multilayer SOM, Fuzzy c-means.


Dimension Reduction with Randomized Anisotropic
Transform for Hyperspectral Image Classification

Huiwu Luo, Lina Yang, Haoliang Yuan and Yuanyan Tang

Department of Computer and Information Science, University of Macau, Macau


Dimension reduction plays an important role in the community of high dimensional data analysis. The notion of ran- dom anisotropic transform (RAT) , which was applied to speed up the computation procedure of dimension reduction kernel(DRK) with Isomap embedding (Isomap-RAT) , was introduced in this paper. Nevertheless, traditional Isomap-RAT does not consider the intrinsic dimension that the hyperspectral image data resides on. Moreover, The DRK of Isomap embedding is not always guaranteed to be positive semi-definite. Thus, this paper proposed a kernel Isomap-Hysime random anisotropic transform (KIH- RAT) to deal with these challenges that met frequently in reality. The proposed methodology consists of two main terms: 1) a kernel term that finds an approximative constant which is added to the dissimilar matrix to make the DRK to be positive semi-definite; and 2) an intrinsic dimension assessment term that employs Hysime to estimate the intrinsic dimension of hyperspectral image data to preserve the geometries of original information as much as possible. The proposed method is exhaustively tested on two reduced feature spaces that relate to the classification of real hyperspectral remote sensing images. The effectiveness and feasibility of presented KIH-RAT methodology are illustrated by the experiment results from both real hyperspectral image examples.

Keywords: Dimension Reduction, Hyperspectral Im- age, Anistropic Transform, Random Projection, Randomized Anisotropic Transform.


Dictionary Learning by Nonnegative Matrix Factorization with L1/2-Norm Sparsity Constraint

Zhenni Li, Zunyi Tang and Shuxue Ding

School of Computer Science and Engineering, The University of Aizu,
Tsuruga, Ikki-Machi, Aizu-Wakamatsu City, Fukushima , Japan


In this paper, we propose an overcomplete, non- negative dictionary learning method for sparse representation of signals, which is based on the nonnegative matrix factorization (NMF) with l1/2-norm as the sparsity constraint. By introducing the l1/2-norm as the sparsity constraint into NMF, we show that the problem can be cast as sequential optimization problems of quadratic functions and quartic functions. The optimization problem of each quadratic function can be solved easily since the problem has closed-form unique solution. The optimization problem of quartic function can also be formulated as solving a cubic equation, which can be efficiently solved by the Cardano formula and selecting one of solutions with a rule. To implement this nonnegative dictionary learning, we develop an algorithm by employing coordinate-wise decent strategy, i.e., coordinate- wise decent based nonnegative dictionary learning (CDNDL). Nu- merical experiments show that the proposed algorithm performs better than the nonnegative K-SVD (NN-KSVD) and the other two compared algorithms.

Keywords: Nonnegative dictionary learning, overcomplete dictionary, sparse representation, NMF.


Sector Projection Fourier Descriptor for Chinese Character Recognition

Dong Li, Jian Wang, Yuanman Li and Yuanyan Tang

Department of Computer and Information Science, University of Macau, Macau, China


Rotation, scale and translation invariant (RST- invariant) feature extraction methods play a critical role in optical Chinese character recognition. However, vast majority of them either has a strict constraint of input image, or has a poor performance on discriminating similar Chinese characters. In this paper, a novel method called Sector Projection Fourier Descriptor (SP-FD) is proposed. SP-FD is a region-based Fourier descriptor which contains two stages. Firstly, the input character image is transformed into polar space through sector-projection, which generates a periodic function named sector-projection shape signature. Secondly, feature vector is obtained through 1-D Fourier transformation on the shape signature. In our method, the input image does not require to be normalized as a precondition, and the shape information of original character is not assumed to be available neither. Moreover, the internal structure of character in the circular direction is exploited. The experimental results show the proposed method can extract the RST-invariant feature effectively, and outperforms the typical algorithms on discriminating similar Chinese characters.

Keywords: Rotation-Scale-Translation invariance, feature extraction, Chinese character recognition, sector projection Fourier descriptor.


Learning Recovered Pattern from Region-Dependent Model for Hyperspectral Imagery

Huiwu Luo, Lina Yang, Haoliang Yuan and Yuanyan Tang

Department of Computer and Information Science, University of Macau, Macau


The Compressive-Projection Principle Component Analysis (CPPCA) technique which recovers hyperspectral image(HSI) data from random projection efficiently, has been proved to be significant in decreasing signal-sensing costs at the sender. Inspired by the fact that the spectral signature of the same ground cover is similar, and two pixels of the neighborhood are likely to belonging to the same ground cover, this paper proposed a novel region-dependent approach CPPCA to recover HSI data. Due to the fact that the region map is critical to our proposed algorithm, herewith we employ a robust supervised Bayesian approach (LORSAL-MLL segmentation) which explores both the spectral and spatial information in an intuitive interpretation with small size samples to segment hyperspectral image into different regions. The CPPCA reconstruction procedure is then employed to each region independently other than each partition individually. The effectiveness and practicability of proposed region-dependent CPPCA (RDCPPCA) reconstructed algorithm is illustrated by real hyperspectral image data set with several criteria measurement.

Keywords: Hyperspectral image segmentation, Hyperspec- tral image reconstruction, Compressive sensing, Principle com- ponent analysis.