2. Amitava Chatterjee, Editor, IEEE Transactions on Vehicular Technology (2013 continuing).

Transcription

1 12.1 SELF APPRAISAL REPORT ON BIOMETRIC AUTHENTICATION SYSTEMS Contributing Faculty Member 1. Amitava Chatterjee (AC), Electrical Engineering Department, Age 45 yrs, Journal Publications 79 (International: 73 (IEEE: 19, Elsevier: 32), National: 06), Edited Book Chapters 05, Conference Publications 30 (International: 18, National: 12), H Index 18 (Source: Google Scholar, last accessed on ), Cumulative Impact Factor , Total Citations 1392 (Source: Google Scholar, last accessed on ), Awarded and ongoing doctoral thesis guidance 08, Awarded and ongoing Masters thesis guidance Special Achievements 1. Amitava Chatterjee, Member, Technical Committee on Imaging Measurements and Systems, IEEE Instrumentation & Measurement Society, USA (November, 2013 continuing). 2. Amitava Chatterjee, Editor, IEEE Transactions on Vehicular Technology (2013 continuing). 3. Amitava Chatterjee, Associate Editor, IEEE Transactions on Vehicular Technology ( ). 4. Amitava Chatterjee, Associate Editor, IEEE Transactions on Instrumentation and Measurement (2013 continuing). 5. Amitava Chatterjee, Associate Editor, IEEE Sensors Journal (2013 continuing). 6. Amitava Chatterjee, Associate Editor, Control Engineering Practice Journal, Elsevier Publishers (2012 continuing). 7. Amitava Chatterjee, Guest Editor, Special Issue on Advances In Evolutionary Optimization Based Image Processing, Engineering Applications of Artificial Intelligence Journal, Elsevier Publishers (Expected Date of Publication: 2014). 8. Amitava Chatterjee, Recipient of Certificate of Appreciation from IEEE Vehicular Technology Society as Top Associate Editor for outstanding contributions in 2012 to IEEE Transactions on Vehicular Technology, Amitava Chatterjee, Visiting Faculty, University of Paris XII, France, as an Enseignant Invité (Invited teacher), May June, Amitava Chatterjee, JSPS (Japan Society for the Promotion of Science) Postdoctoral Fellowship (FY ), University of Electro communications, Tokyo, Japan, November, 2004 to November, Amitava Chatterjee, Visiting Faculty, University of Paris XII, France, as an Enseignant Invité (Invited teacher), March May, Amitava Chatterjee, Japanese Government Scholarship (Monbukagakusho), Saga University, Japan, Amitava Chatterjee, Best Paper Presentation Award, 26 th National Systems Conference (NSC 2002), in the area of Cognitive Systems and Artificial Intelligence, INCOIS, Hyderabad, India, November, Relevant Projects in Last 10 years including the Ongoing Projects

2 Project Title Sponsoring Agency Members Grant Value (Rs in Lakh) Duration Solving Simultaneous Localization and Mapping Problems of Mobile Robots by Adaptive Extended Kalman Filtering AICTE, India AC (PI), M Maitra Vision Based Autonomous Navigation of Mobile Robots UGC, India AC (PI), S Munshi, BBhattacharyya Relevant Publications in Last 5 years a) Books: 1. Amitava Chatterjee, Anjan Rakshit, and N. Nirmal Singh, Vision Based Autonomous Robot Navigation: Algorithms And Implementations, Springer Verlag GmbH Heidelberg, Germany, Publishers, ISBN: (Print), (Online) 2. Amitava Chatterjee and Patrick Siarry (Eds.), Computational Intelligence in Image Processing, Springer Verlag GmbH Heidelberg, Germany, Publishers, ISBN: (Print), (Online) 3. Amitava Chatterjee, Hadi Nobahari, and Patrick Siarry (Eds.), Advances in Heuristic Signal Processing and Applications, Springer Verlag GmbH Heidelberg, Germany, Publishers, ISBN: (Print), (Online) b) International Journals: 1. Soumyasree Chakraborty, Indrani Bhattacharya, and Amitava Chatterjee, A palmprint based biometric authentication system using dual tree complex wavelet transform, Measurement (Impact Factor: 1.130), Elsevier publishers, vol. 46, issue 10, pp , December Amitava Chatterjee, Régis Fournier, Amine Naït Ali, and Patrick Siarry, A postural information based biometric authentication system employing S Transform, radial basis function network and extended Kalman filtering, IEEE Transactions on Instrumentation and Measurement (Impact Factor: 1.357), vol. 59, issue 12, pp , December DOI: /TIM Ankush Chakrabarty, Harsh Jain, and Amitava Chatterjee, Volterra kernel based face recognition using artificial bee colony optimization, Engineering Applications of Artificial Intelligence (Impact Factor: 1.625), Elsevier publishers, vol. 26, issue 3, pp , March Kaushik Das Sharma, Amitava Chatterjee, and Anjan Rakshit, Harmony search based hybrid stable adaptive fuzzy tracking controllers for vision based mobile robot navigation, Accepted for publication in Machine Vision and Applications (Impact Factor: 1.103), Springer publishers. DOI: /s z 5. Amitava Chatterjee, Patrick Siarry, Amir Nakib, and Raphael Blanc, An improved biogeography based optimization approach for segmentation of human head ct scan images employing fuzzy entropy, Engineering

3 Applications of Artificial Intelligence (Impact Factor: 1.625), Elsevier publishers, vol. 25, pp , December Anjan Rakshit and Amitava Chatterjee, A microcontroller based compensated optical proximity detector employing switching mode synchronous detection technique, Measurement Science and Technology (Impact Factor: 1.435), IOP Science Publishers, vol. 23, no. 3, March doi: / /23/3/ Kaushik Das Sharma, Amitava Chatterjee, and Anjan Rakshit, A PSO Lyapunov hybrid stable adaptive fuzzy tracking control approach for vision based robot navigation, IEEE Transactions on Instrumentation and Measurement (Impact Factor: 1.357), vol. 61, no. 7, pp , July DOI: /TIM Nandita Sanyal, Amitava Chatterjee, and Sugata Munshi, An adaptive bacterial foraging algorithm for fuzzy entropy based image segmentation, Expert Systems with Applications (Impact Factor: 1.854), Elsevier Publishers, vol. 38, issue 12, pp , November December Avishek Chatterjee, Olive Ray, Amitava Chatterjee, and Anjan Rakshit, Development of a real life EKF based SLAM system for mobile robots employing vision sensing, Expert Systems with Applications (Impact Factor: 1.854), Elsevier Publishers, vol. 38, issue 7, pp , July Avishek Chatterjee, N. Nirmal Singh, Olive Ray, Amitava Chatterjee, and Anjan Rakshit, A two camera based vision system for image feature identification, feature tracking and distance measurement by a mobile robot, International Journal of Intelligent Defence Support Systems, Inderscience Publishers, vol. 4, no. 4, pp , /IJIDSS N. Nirmal Singh, Avishek Chatterjee, Amitava Chatterjee, and Anjan Rakshit, A two layered subgoal based mobile robot navigation algorithm with vision system and IR sensors, Measurement (Impact Factor: 1.130), Elsevier Publishers, vol. 44, issue 4, pp , May Akhilesh Chander, Amitava Chatterjee, and Patrick Siarry, A new social and momentum component adaptive PSO algorithm for image segmentation, Expert Systems with Applications (Impact Factor: 1.854), Elsevier Publishers, vol. 38, issue 5, pp , May Susanta Ray, Gautam Sarkar, Amitava Chatterjee, and Anjan Rakshit, Development of a microcontroller based frequency synthesizer cum vector voltmeter, IEEE Sensors Journal (Impact Factor: 1.475), vol. 11, no. 4, pp , April DOI: /JSEN Anjan Rakshit and Amitava Chatterjee, A microcontroller based IR range finder system with dynamic range enhancement, IEEE Sensors Journal (Impact Factor: 1.475), vol. 10, no. 10, pp , October DOI: /JSEN N. Nirmal Singh, Amitava Chatterjee, and Anjan Rakshit, A PIC microcontroller system for real life interfacing of external peripherals with a mobile robot, International Journal of Electronics (Impact Factor: 0.440), Taylor and Francis Publishers, vol. 97, issue 2, pp , DOI: /

4 16. Madhubanti Maitra and Amitava Chatterjee, A novel technique for multilevel optimal magnetic resonance brain image thresholding using bacterial foraging, Measurement (Impact Factor: 1.130), Elsevier Publishers, vol. 41, issue 10, pp , December Madhubanti Maitra and Amitava Chatterjee, Hybrid multiresolution Slantlet transform and fuzzy c means clustering approach for normal pathological brain MR image segregation, Medical Engineering & Physics (Impact Factor: 1.779), Elsevier Publishers, vol. 30, issue 5, pp , June DOI: /j.medengphy Madhubanti Maitra and Amitava Chatterjee, A hybrid cooperative comprehensive learning based PSO algorithm for image segmentation using multilevel thresholding, Expert Systems with Applications (Impact Factor: 1.854), Elsevier Publishers, vol. 34, issue 2, pp , February Madhubanti Maitra and Amitava Chatterjee, A Slantlet transform based intelligent system for magnetic resonance brain image classification, Biomedical Signal Processing and Control (Impact Factor: 1.074), Elsevier Publishers, vol. 1, issue 4, pp , Oct Note: Detail list of publications of Dr. Amitava Chatterjee is available at: HOMEPAGE: GOOGLE SCHOLAR CITATION PAGE: Facilities Available 1. SONEL S. A. make KT 384 Thermal Imager, manufactured by SONEL, Swidnica, Poland., procured utilizing the financial support received from DST PURSE Programme Ongoing Work under UPE II The general methodology followed for development of such systems involves the steps of a) Sensory Data Acquisition b) Preprocessing and Feature Extraction c) Classification/Verification System d) Automated Decision Making The following solutions have been developed under the three broad categories of BASs considered in this project: Category I: Palm Dorsa Vein Pattern (PDVP) based solutions using Thermal Imaging. 1. Title: Multi Size Patch based Collaborative Representation for Palm Dorsa Vein Pattern Recognition by Enhanced Ensemble Learning with Modified Interactive Artificial Bee Colony Algorithm Authors: Sandip Joardar, Amitava Chatterjee, Sanghamitra Bandyopadhyay, and Ujjwal Maulik Journal: IEEE Transactions on Systems, Man and Cybernetics: Systems Abstract: This work proposes a novel method, a Collaborative Representation based Classification (CRC) strategy, for palm dorsa vein pattern (PDVP) based human recognition employing thermal imaging. This thermal PDVP

5 imaging based human recognition methodology has been specifically employed to encounter the challenging crisis of intrusion posed by imposters. To address the Small Sample Size (SSS) problem, intrinsic to many biometric applications, each image is transformed into multiple patches, leading to an increase in the total number of samples. In a bid to make the classification strategy less sensitive to the choice of patch size, the present paper proposes a new enhanced ensemble learning for the patch based CRC (PCRC) algorithm, where margin maximization is carried out using exponential squared loss minimization. This work also proposes how this loss minimization can be achieved by a stochastic optimization algorithm and solves this problem using artificial bee colony (ABC) algorithm. In this context, a new ABC variant, called modified interactive artificial bee colony (MI ABC) algorithm, has also been proposed, which has been demonstrated to outperform the basic ABC and its several modern variants. The proposed methodology has been implemented on a well structured real database, developed in our laboratory using real subjects, and the results obtained in implementation phase clearly demonstrate that our proposed method could outperform its several competitors and achieve substantially high classification rates, for the problem under consideration. Status: Submitted for First Round of Technical Review Figure and Table pertaining to this study Fig. 1. Data acquisition, signal conditioning and image preprocessing steps employed in our laboratory for FIR imaging of PDVP. TABLE 1: Biometric Recognition Performance Comparison METHOD AVERAGE RECOGNITION RATE (%) MSPCRC MSPCRC with modified CRC_RLS EMSPCRC with MI ABC TRIAL 1 TRIAL 2 TRIAL 3 Top Recognition Rate Worst Recognition Rate

6 2. Title: Palm Dorsa Vein Pattern Recognition by Autonomous Collaborative Representation Based Classification Using Lanczos Bidiagonalization Authors: Sandip Joardar and Amitava Chatterjee Journal: IEEE Transactions on Information Forensics and Security Abstract: This work proposes a new Collaborative Representation (CR) based Palm Dorsa Vein Pattern (PDVP) classification strategy for human recognition using Infrared imaging based thermal images of PDVP. The training dictionary that is formed using the thermal images of the PDVP is, generally, ill conditioned in nature, posing an ill posed problem to the classification strategy. Most of the existing algorithms of Sparse and Collaborative Representation based Classification fail to perform under those circumstances where one needs to perform such human recognition with an ill conditioned training dictionary, under the challenging training situation of Small Sample Size (SSS) scenario. The present research proposes an improved classification strategy, especially for those challenging situations, named as Autonomous Collaborative Representation based Classification (Auto_CRC). The Auto_CRC algorithm proposes to solve such ill posed problems by using Lanczos Bidiagonalization method, wherein we propose to utilize the L curve method for suitable determination of the regularization parameter. The proposed method is implemented on a well structured database, JU TV1: Thermal Vein Database, which has been developed in the Electrical Instrumentation and Measurement Laboratory of Electrical Engineering Department, Jadavpur University, Kolkata, India, using real subjects. The results obtained sufficiently demonstrate the substantially high and stable recognition rates achievable in the final classification utilizing the proposed method. The results too depict the ease with which the proposed method outperforms several of its well known, and extensively implemented, competitors. Status: Submitted for First Round of Technical Review Figure and Table pertaining to this study Same Fig. 1, as mentioned in the previous study, is used here. TABLE 2: Performance comparison for three competing methods using L Curve METHOD AVERAGE RECOGNITION RATE (%) BEST RESULT (%) Standard SRC 1 training sample = training samples = training samples = training samples = CRC_RLS 1 training sample= training samples = training samples = training samples = Auto_CRC(proposed) 1 training sample = training samples = training samples = training samples = Category II: Palm Print Recognition based BAS

7 1. Title: A Conglomeration of Cross Wavelet Transform and Bacterial Foraging Optimized Feature Selection For Palm Print Authentication Authors: Nandita Sanyal, Amitava Chatterjee, and Sugata Munshi Journal: Measurement (Elsevier) Abstract: The work puts forth a new palm print based biometric authentication system. It is low cost, as it employs a scanner coupled to a PC, for acquiring the image of the human palm. The histogram of the palm image is cross wavelet transformed with respect to that of a reference palm image. Several features are extracted from the resulting spectrum, and a trained ANN is used to identify the subject, based on these features. Classical Bacterial Foraging Optimization Algorithm is used to select the combination of features that results in the best performance. An Adaptive Bacterial Foraging Optimization Algorithm is also introduced to further augment the performance of the system. Extensive experimental investigations reveal that an authentication accuracy of more than 97% can be obtained with the system. Status: Submitted for First Round of Technical Review Figure and Table pertaining to this study TABLE 3: OVERALL AVERAGE CLASSIFICATION PERFORMANCE OF PROPOSED BIOMETRIC AUTHENTICATION SYSTEMS WITH AND WITHOUT FEATURE SELECTION With 12 number of total features Overall classification efficiency (%) With 14 number of total features Overall classification efficiency (%) With 16 number of total features Overall classification efficiency (%) Without feature Feature by CBFOA Feature by ABFOA Without feature Feature by CBFOA Feature by ABFOA Without feature Feature by CBFOA Feature by ABFOA Original palm print Gray palm print Gray palm image after cropping w.r.t. largest palm print Corresponding Histogram Subject 1 Subject 2 Fig. 4. Acquired palm print, colored image to gray conversion, cropped gray palm print and corresponding histogram for palm images of sample subjects.

8 Category III: Face Recognition based BAS. 1. Title: A novel binary adaptive weight GSA based feature for face recognition using local gradient patterns, modified census transform, and local binary patterns Authors: Tapabrata Chakraborti and Amitava Chatterjee Journal: Engineering Applications of Artificial Intelligence (Elsevier) Abstract: The present work proposes a novel feature scheme for face recognition problems, employing a new modified version of the gravitational search algorithm, a recently proposed metaheuristic optimization algorithm. The feature scheme, which reduces the dimensionality of the set of extracted features by choosing the features with high discriminative power, has been employed in conjunction with three contemporary feature extraction algorithms popularly employed for face recognition purposes, namely local binary pattern (LBP), modified census transform (MCT), and local gradient pattern (LGP) algorithms. The feature is carried out by formulating a fitness function as a ratio of the within class distance to the between class distance and then a binary version of traditional GSA is developed for solving this problem. This binary GSA (named BGSA) is further enhanced to propose a novel binary variation of GSA with dynamic adaptive inertia weight (named BAW GSA). Six new algorithms for face recognition are proposed hybridizing BGSA or BAW GSA with each of LBP, MCT and LGP algorithms. In each algorithm, the classification step is carried out using backpropagation neural network. The algorithms were extensively tested for five benchmark face databases (Yale A, Yale B extended, ORL, LFW and AR) and it was conclusively proven that our proposed LGP hybridized with BAW GSA algorithm comfortably outperforms several competing, contemporary algorithms existing in literature. Status: Revised version of the manuscript under preparation (original review decision received: Major Revision) Table 4: Face recognition percentage error rates for Yale A database Method Training size S LDA S LDA (updated) TANMM OLAP Eigenfaces Fisherfaces Laplacianfaces Volterrafaces(linear) Volterrafaces(quadratic) LBP MCT LGP LBP + BGSA MCT + BGSA LGP + BGSA LBP + BAW GSA MCT + BAW GSA LGP + BAW GSA

9 Table 5: Face recognition percentage error rates for Yale B extended database Method Training size ORO 9.0 SR RDA KLPPSI KRR CTA Eigenfaces Fisherfaces Laplacianfaces MLASSO Volterrafaces(linear) Volterrafaces (quadratic) LBP MCT LGP LBP + BGSA MCT + BGSA LGP + BGSA LBP + BAW GSA MCT + BAW GSA LGP + BAW GSA Title: A Novel Local Extrema Based Gravitational Search Algorithm and its Application in Face Recognition using One Training Image per Class Authors: Tapabrata Chakraborti, Kaushik Das Sharma, and Amitava Chatterjee Journal: Engineering Applications of Artificial Intelligence (Elsevier) Abstract: In this present work a new methodology has been presented involving a stochastic optimization based approach to solve the face recognition problem with only one training image per class. Singular value decomposition (SVD) is used to decompose the single training image into two component images in order to compute the within class scatter matrix. The stochastic optimization approach is implemented employing gravitational search algorithm (GSA) which searches for an optimal transform matrix instead of using the traditional solution of general eigenvalue problem as is carried out in Fisher linear discriminant analysis (FLDA). The present paper also proposes two novel variants of GSA, namely the 2 D version of GSA, in order to cater for the 2 D image data, and the other one is a 2 D randomized local extrema based GSA (RLEGSA), which employs a stochastic local neighbourhood based search instead of global search, as in basic GSA. Finally, a novel concept of performing an automated of projection vectors is incorporated in the 2 D RLEGSA to propose an improved variant, called the Modified RLEGSA (MRLEGSA). Experimental results, based on benchmark Yale A and ORL databases, show that the proposed methods outperform several existing schemes.

10 Status: Revised version of the manuscript submitted (original review decision received: Major Revision) Figures pertaining to this study Fig. 5. A comparison of recognition accuracy vs. number of projection vectors for five competing methods, for Yale A dataset. Fig. 6. A comparison of recognition accuracy vs. number of projection vectors for five competing methods, for ORL dataset.

11 3. Title: Collaborative representation based face recognition using a hybrid similarity measure with single training sample per person Authors: Sandip Joardar and Amitava Chatterjee Conference: International Conference on Control, Instrumentation, Energy & Communication (CIEC14), Kolkata, India, January 31 February 02, Abstract: This work proposes a Collaborative Representation (CR) based face recognition strategy, wherein a hybrid similarity measure has been used for the computation of the reconstruction residual. In this work Single Sample per Person (SSPP) methodology is used for creating the training dataset. The hybrid similarity measure proposed for computation of the reconstruction residual is a composite formulation of the Procrustes Similarity analysis along with the conventional Euclidean Distance metric. The proposed method is implemented for several benchmark face databases and the results obtained demonstrate the substantial improvement achieved in the recognition rates, in comparison to the known Collaborative Representation (CR) based face recognition strategy. Status: Accepted for presentation Tables pertaining to this study Table 6: Performance comparison for AT & T database AVERAGE RECOGNITION RATE (%) METHOD TRIAL 1 TRIAL 2 TRIAL 3 CRC_RLS ± ± ± CRC_HSM ± ± ± Table 7: Performance comparison for Yale A database METHOD AVERAGE RECOGNITION RATE (%) TRIAL 1 TRIAL 2 TRIAL 3 CRC_RLS ± ± ± CRC_HSM ± ± ± Table 8: Performance comparison for Sheffield database METHOD AVERAGE RECOGNITION RATE (%) TRIAL 1 TRIAL 2 TRIAL 3 CRC_RLS ± ± ± CRC_HSM ± ± ±

12 Most significant conclusions 1. A comprehensive Palm Dorsa Vein Pattern (PDVP) thermal image database has been prepared using real subjects, named as JU TV1: Thermal Vein Database, which has been developed in the Electrical Instrumentation and Measurement Laboratory of Electrical Engineering Department, Jadavpur University, Kolkata, India. It is hoped that this database will serve as a benchmark database for similar researches carried out in this domain, all over the world. At present, significant revamping of the database is under progress to include more number of subjects under varied conditions, to make the database more comprehensive in nature. 2. Two solutions have been developed for PDVP based biometric authentication systems, both for the challenging scenario of Small Sample Size (SSS) problems. In one solution, multi size patch based Collaborative Representation for PDVP recognition by Enhanced Ensemble Learning has been implemented using Modified Interactive Artificial Bee Colony algorithm. In the other solution, an Autonomous Collaborative Representation based Classification algorithm is developed using Lanczos Bidiagonalization method, wherein the L curve method is used to determine the regularization parameter. Both methods could achieve encouraging mean recognition accuracies of 94 97%. 3. A new low cost solution for palm print based biometric authentication systems has been proposed utilizing histogram of each image, and then using Cross Wavelet Transform based feature extraction and ANN based authentication. An Adaptive Bacterial Foraging Optimization algorithm based solution is developed for suitable feature purpose and the system could produce 97% recognition accuracy for the palm print database indigenously developed in our laboratory. 4. In face recognition based biometric systems, two schemes have been developed using Gravitational Search Algorithm (GSA). In one scheme, a novel feature method is employed using a modified GSA which is hybridized with LBP, MCT, and LGP algorithms. In the other scheme, GSA has been hybridized with SVD to solve the challenging problem of Single training Image per Person (SIPP). Both algorithms achieved excellent recognition performances for several benchmark face databases. 5. Another scheme for face recognition has been proposed for the SIPP problem using Collaborative Representation based Classification and a hybrid similarity measure, using Procrustes Similarity analysis. This scheme has also achieved encouraging performance for three benchmark face databases.