International Journal of Biometrics (8 papers in press)
Signature Recognition Using Binary Features and KNN
by Hedjaz Hezil, Rafik Djemili, Houcine Bourouba
Abstract: This paper proposes the use of binary features in offline signature recognition systems. Indeed, offline signature recognition finds mainly its importance for the authentication of administrative and official documents in which a higher accuracy is needed. In the proposed approach, features are extracted by using two descriptors: Binary Statistical Image Features (BSIF) and Local Binary Patterns (LBP). To assess the reliability of the method, experiments were carried out using two publicly available datasets, MCYT-75 and GPDS-100 databases. Using a k-nearest neighbor classifier, recognition performances reach values high as 97.3% and 96.1% for MCYT-75 and GPDS-100 databases respectively.In signature verification, the classification accuracy measured with equal error rate (EER) achieved 4.2% and 4.8% respectively on GPDS-100 and GPDS-160. In addition, the EER for the MCYT-75 database has attained 7.78%. All those accuracies outperformed various performance results reported in literature.
Keywords: Offline signature recognition; feature extraction; biometric; LBP; BSIF; KNN.
Palmprint Identification and Verification with minimal number of features
by Hemantha Kumar Kalluri
Abstract: In this paper, palmprint verification and identification with minimum number of features is proposed. Apply the Wide Principal Line Extractors (WPLEs) on the Region of Interest (ROI) to generate Wide Principal Line Images (WPLIs). The WPLI is segmented into 2x2, 4x4, 8x8 and 16x16 and the feature value is extracted directly from each segment. Experiments are conducted by using the extracted features. The results show that the Equal Error Rate (EER), Decidability Index (DI) and Correct
Recognition Rate (CRR) of the proposed approach is better than existing methods for PolyUPalmprint Database.
Keywords: Palmprint; Biometrics; Region of Interest; Feature Extraction; PalmprintrnIdentification; Palmprint Verification.
GSI: Efficient Spatio-Temporal Template for Human Gait Recognition
by Mohammad H. Ghaeminia, Shahriar B. Shokouhi
Abstract: Human gait recognition is a challenging task in computer vision community. In order to represent the gait in video sequences, the most common feature is a gait template. Many efficient templates have been developed recently, however, the effectiveness of the proposed gait motion models is still under investigation. A novel template-based feature, named Gait Salient Image (GSI) is introduced in this paper. The main contribution of the proposed GSI template is encoding the motion energy of gait into a single template. This idea is being conceptualized by applying appropriate spatio-temporal filter for extracting motion features from the sequences and averaging it over a gait period. To show how GSI-based feature is being efficient, the proposed template is classified using PCA+LDA. Extensive experiments on popular gait databases reveal an improvement over the available methods in terms of efficiency and accuracy. The value of recognition rate is 58.44% for Rank1 and 76.60% for Rank5 based on the USF database.
Keywords: gait recognition; spatio-temporal filtering; template-based features.
On the Potential of EEG for Biometrics: Combining Power Spectral Density with a Statistical Test
by Hemang Shrivastava, Gleb Tcheslavski
Abstract: The objective of this work was to explore the potential of using subjects electroencephalogram (EEG) as a bio-metric identifier. EEG was collected from eight healthy male participants, while exposing them to the sequence of images displayed on the screen. The averaged, over EEG rhythms, Power Spectral Density estimates were used as the classification features for the Artificial Neural Network and Euclidean Distance-based classifiers. Prior the classification, Kruskal-Wallis test was performed on the power estimates to verify that they were statistically different between different individuals, who were performing identical tasks. Assuming the significance level of 0.075, Kruskal-Wallis analysis indicated that up to 96.42% of such estimates were statistically different between different participants and, therefore, can be used as the classification features for bio-metric authentication. When using average EEG spectral power as the classification features, the highest classification accuracy of 87.5% was achieved for low alpha EEG rhythm (810 Hz), while using the Artificial Neural Network classifier, and for high alpha EEG rhythm (1014 Hz), while using the Euclidean Distance classifier. The classification performance may be mediated by the type of visual stimulation (i.e., the image the subject perceives) and the statistical test may be instrumental for classification feature selection.
Keywords: Biometric authentication; Electroencephalogram (EEG); Brain-Computer Interface; Bioinformatics.
An Authentication System using Keystroke Dynamics
by Farhana Javed Zareen, Chirag Matta, Akshay Arora, Sarmod Singh, Suraiya Jabin
Abstract: There are various biometrics based methods for user authentication. However, best authentication method can be based on physiological/behavioural biometrics as capturing physiological biometrics may require use of special devices and that may not be available with many users. Keystroke dynamics is a simplified and easily achievable user authentication method when every user is available with a laptop or a personal computer. This paper presents a keystroke dynamics based authentication system using Bayesian regularized feed-forward neural network. In order to train the model, a database is captured for recording keystroke dynamics of twenty users in four sessions each with fifty samples. Experimental results demonstrate that the Bayesian regularized neural network models provide the best results and are most suitable for this purpose. We are able to achieve an equal error rate of 0.9% which is better than the methods used in the existing literature for plain keystroke dynamics. We have given a comparative analysis of the performance of proposed system with existing methods.
Keywords: individual authentication; biometrics; equal error rate; keystroke dynamics; pattern recognition; machine learning.
A novel fingerprint classification system using BPNN with local binary pattern and weighted PCA
by Sasirekha Kathirvel
Abstract: Fingerprint classification is an important indexing scheme to reduce fingerprint matching time for large volumes of the database. In this paper, a novel approach to classify the fingerprint images is proposed. It involves four main parts: Denoising, feature extraction, dimensionality reduction and classification. Initially, the fingerprint image is denoised using undecimated wavelet transform with a threshold based on the golden ratio and weighted median. Then Short Time Fourier Transform (STFT) is used to enhance the denoised fingerprint image. Conventional fingerprint classification algorithms use the core and delta points to classify fingerprint images into common classes and require that these points be present in the image. They may not exploit the rich discriminatory information existing in the fingerprints. The detection of core and delta points is very difficult due to the presence of creases, scars, smudges, dryness or blurred prints. It is proposed to extract Local Binary Pattern (LBP) features from the fingerprint to overcome the difficulty associated with singular point detection. To reduce the dimensionality of the feature space Quick Reduct (QR), Principal Component Analysis (PCA) and the proposed weighted PCA have been investigated.Then use these features to train a Back Propagation Neural Network (BPNN) for classifying fingerprints. In this research, experiments have been conducted on real-time fingerprint images collected from 150 subjects each with ten samples using eSSL ZK7500 scanner and also on the NIST-4 dataset. To validate the performance of the proposed fingerprint classification method different measures such as accuracy, sensitivity, precision, recall and error rate have been computed and compared with other benchmark classification techniques such as Support Vector Machine (SVM), K-Nearest Neighbor (K-NN), and Multi-Layer Perceptron (MLP).
Keywords: Fingerprint; classification; LBP; BPNN; Quick Reduct; weighted PCA.
Recognition of Ear based on Partial Features Fusion
by Vidyasri Ramesh, Priyalakshmi B, Ravi Raja M, Priyanka S
Abstract: Multi biometric systems like face and ear recognition techniques are adopted for the forensic and civilian applications to address the challenges of the facial expressions and occlusions. Numerous face and ear techniques have been proposed so far. Yet it becomes difficult to remove occlusions in ear. Ear occlusions can be of various forms such as, cap, hair, scarf, earrings, etc. Due to occlusion during Identification stage recognition will certainly cause the loss of Information. In this paper, an Occlusion detection of Ear that will recognize the occlusion information during the Identification stage and by using fusion method, the matching of the samples are processed.
Keywords: Ear recognition; occlusion estimation; biometric system; FLDA feature extraction; borda count; rank features.
Face Analysis in Video : Face Detection and Tracking with Pose Estimation
by Hazar Mliki, Mohamed Hammami
Abstract: We introduced a full automatic approach to achieve face detection and tracking with pose estimation in video sequences. The proposed approach consists of three modules: face detection module, face tracking module and face pose estimation module. A combination between detection and tracking modules was performed to overcome the different challenging problems that might occur while detecting or tracking faces. Afterward face pose estimation module was applied to select the best camera capture which is closest to the frontal face view for better face recognition task. The performance of these modules was evaluated with an experimental study which has proven the robustness of the proposed approach for a face analysis in video.
Keywords: Face Detection; Face Tracking; Face Pose Estimation; Data-Mining; SVM; Adaboost.