Score Level Fusion Technique Research Articles

Domain Adaptation for Unconstrained Ear Recognition with Convolutional Neural Networks

Ear recognition has gained attention within the biometrics community recently. Ear images can be captured from a distance without contact, and the explicit cooperation of the subject is not needed. In addition, ears do not suffer extreme change over time and are not affected by facial expressions. All these characteristics are convenient when implementing surveillance and security applications. At the same time, applying any Deep Learning (DL) algorithm usually demands large amounts of samples to train networks. Thus, we introduce a large-scale database and explore fine-tuning pre-trained Convolutional Neural Networks (CNN) to adapt ear domain images taken under uncontrolled conditions. We built an ear dataset from the VGGFace dataset by profiting the face recognition field. Moreover, according to our experiments, adapting the VGGFace model to the ear domain leads to a better performance than using a model trained on general image recognition. The efficiency of the trained models has been tested on the UERC dataset achieving a significant improvement of around 9\% compared to approaches in the literature. Additionally, a score-level fusion technique was explored by combining the matching scores of two models. This fusion resulted in an improvement of around 4\% more. Open-set and close-set experiments have been performed and evaluated using Rank-1 and Rank-5 recognition rate metrics

Real-Time Hand Gesture Recognition Using Fine-Tuned Convolutional Neural Network.

Hand gesture recognition is one of the most effective modes of interaction between humans and computers due to being highly flexible and user-friendly. A real-time hand gesture recognition system should aim to develop a user-independent interface with high recognition performance. Nowadays, convolutional neural networks (CNNs) show high recognition rates in image classification problems. Due to the unavailability of large labeled image samples in static hand gesture images, it is a challenging task to train deep CNN networks such as AlexNet, VGG-16 and ResNet from scratch. Therefore, inspired by CNN performance, an end-to-end fine-tuning method of a pre-trained CNN model with score-level fusion technique is proposed here to recognize hand gestures in a dataset with a low number of gesture images. The effectiveness of the proposed technique is evaluated using leave-one-subject-out cross-validation (LOO CV) and regular CV tests on two benchmark datasets. A real-time American sign language (ASL) recognition system is developed and tested using the proposed technique.

Facial Expression Recognition Using Local Gravitational Force Descriptor-Based Deep Convolution Neural Networks

An image is worth a thousand words; hence, a face image illustrates extensive details about the specification, gender, age, and emotional states of mind. Facial expressions play an important role in community-based interactions and are often used in the behavioral analysis of emotions. Recognition of automatic facial expressions from a facial image is a challenging task in the computer vision community and admits a large set of applications, such as driver safety, human-computer interactions, health care, behavioral science, video conferencing, cognitive science, and others. In this work, a deep-learning-based scheme is proposed for identifying the facial expression of a person. The proposed method consists of two parts. The former one finds out local features from face images using a local gravitational force descriptor, while, in the latter part, the descriptor is fed into a novel deep convolution neural network (DCNN) model. The proposed DCNN has two branches. The first branch explores geometric features, such as edges, curves, and lines, whereas holistic features are extracted by the second branch. Finally, the score-level fusion technique is adopted to compute the final classification score. The proposed method along with 25 state-of-the-art methods is implemented on five benchmark available databases, namely, Facial Expression Recognition 2013, Japanese Female Facial Expressions, Extended CohnKanade, Karolinska Directed Emotional Faces, and Real-world Affective Faces. The databases consist of seven basic emotions: neutral, happiness, anger, sadness, fear, disgust, and surprise. The proposed method is compared with existing approaches using four evaluation metrics, namely, accuracy, precision, recall, and f1-score. The obtained results demonstrate that the proposed method outperforms all state-of-the-art methods on all the databases.

Score Level Fusion Technique for Human Identification

A multimodal biometric-system based score level fusion technique is proposed to construct a robust human identification system. Feature fusion can be implemented via different methods. In this paper, the score level fusion of face and iris traits are combined and re-classified at Equal Error Rate (EER) value to improve the individual unimodal systems performance for recognizing 80 subjects (40 subject per one face-iris dataset). The multimodal classification results are compared and evaluated comprehensively using four competitive feature extraction methods: Principle Component Analysis (PCA), Fourier Descriptors (FDs), Gray Level Co-occurrence Matrix (GLCM), and Local Binary Pattern (LBP). Also, a low-quality resolution of MMU1 iris database are considered in this work as an additional challenge for system robustness. The accuracy rate of GLCM and LBP methods satisfied 100% with ORL-CASIA-V1 combination datasets, while PCA and GLCM methods achieved 100% with the low-quality ORL-MMU-1 combination datasets, these results provide an evidence of how the multimodal biometric system could improve the overall unimodal systems performance. Also, the GLCM advances all other feature extraction methods by having the highest accuracy rate with ORL-CASIA-V1 and ORL-MMU-1 combined datasets.

Presentation attack detection based on score level fusion and challenge-response technique

Biometrics is the state of the art in dealing with identity identification and verification based on the physical and behavioral characteristics and widely used in the fields of Fintech, such as mobile payment and online banking due to its security and convenience. However, there are various attacks against the biometrics system. The presentation attack is one of the most common attacks that an imposter presents fake biometrics to the sensor trying to fool the system. This paper proposes a multimodal presentation attack detection (PAD) method against photo-attack and video-attack in face recognition system by using score level fusion and challenge-response scenario. The proposed challenge-response scenario is that requesting the user to speak out the randomly prompted words. Then, the recognized speech text and the user’s mouth motion are detected simultaneously to verify if the user is liveness. Two weighted score level fusion rules, namely weighted sum and weight product, are used to combine the speech and mouth motion traits as a matching score. The final score is fed into supervised machine learning algorithms and trained for classifying spoofing. The experiments are conducted in the self-built database. Experimental results show that the proposed method can achieve the best half total error rate at 3.64% and can effectively improve facial recognition system security.

Ear anti-spoofing against print attacks using three-level fusion of image quality measures

Ear biometric recognition systems have been developed in the last decades. Consequently, the attacks to these biometrics security systems became inevitable. Recently, image quality assessment methods are one of the popular methods used for countering spoof attacks. There is no ear anti-spoofing system proposed in the literature. Therefore, in order to counter spoof attacks on ear biometric recognition systems, an efficient method is proposed in this study. This proposed method is based on full-reference and no-reference image quality assessment methods. Three-level score-level fusion and decision-level fusion techniques are employed in the solution of the proposed method. The experiments are conducted by preparing the printed photo attack images of AMI and UBEAR ear databases. The proposed system significantly recognizes real and fake ear images compared to the other systems implemented in this study. Half total error rates of the proposed system using printed photo attack images are compared with the error rates of the other implemented systems employing various fusion techniques for ear anti-spoofing. Additionally, the proposed system is compared with the state-of-the-art anti-spoofing systems and CNN-based deep learning anti-spoofing systems against print attacks on other biometric traits since this is the first study presenting ear anti-spoofing systems using score-level and decision-level fusion of image quality assessment methods.

Integrating Feature Extractors for the Estimation of Human Facial Age

ABSTRACTFacial feature extraction algorithms play an important role in many applications of face biometrics such as face recognition for person identification, classification of emotions by facial expression recognition and age estimation using facial images. In this paper, an integration of different type of feature extraction algorithms is applied on facial images for accurate age estimation. This integration is performed by using two-level fusion of features and scores with the help of feature-level and score-level fusion techniques. In our proposed method, the advantage of using different types of features such as biologically inspired features, texture-based features, and appearance-based features is used. Feature-level fusion of biologically inspired and texture-based methods is integrated into the proposed method and their combination is fused with an appearance-based method using score-level fusion. Experiments on the publicly available MORPH and FG-NET databases prove the effectiveness of the proposed method and the proposed method outperforms many of the state-of-the-art systems.

I-Vector-Based Speaker Verification on Limited Data Using Fusion Techniques

Abstract In many biometric applications, limited data speaker verification plays a significant role in practical-oriented systems to verify the speaker. The performance of the speaker verification system needs to be improved by applying suitable techniques to limited data condition. The limited data represent both train and test data duration in terms of few seconds. This article shows the importance of the speaker verification system under limited data condition using feature- and score-level fusion techniques. The baseline speaker verification system uses vocal tract features like mel-frequency cepstral coefficients, linear predictive cepstral coefficients and excitation source features like linear prediction residual and linear prediction residual phase as features along with i-vector modeling techniques using the NIST 2003 data set. In feature-level fusion, the vocal tract features are fused with excitation source features. As a result, on average, equal error rate (EER) is approximately equal to 4% compared to individual feature performance. Further in this work, two different types of score-level fusion are demonstrated. In the first case, fusing the scores of vocal tract features and excitation source features at score-level-maintaining modeling technique remains the same, which provides an average reduction approximately equal to 2% EER compared to feature-level fusion performance. In the second case, scores of the different modeling techniques are combined, which has resulted in EER reduction approximately equal to 4.5% compared with score-level fusion of different features.

MULTI-BIOMETRIC AUTHENTICATION SYSTEM USING FEATURE LEVEL FUSION

Advancements in science, technologies, industries and in numerous other fields have lead to the need for security improvements in order keep their valuables, workplace, societies, homes, etc., accessible only to authorized personnel. Biometric authentication is a well-known solution for this issue. One of the most unique patters can be obtained in finger veins which is also a unique physiological biometric. And since finger veins are beneath the skin, they are considered to be much reliable, secure and also know to be forge proof. Finger vein authentication system has proven to be a remarkable technology in the field of biometrics and are applicable where higher levels of security and privacy is very important. This paper, proposes a feature level fusion based system that uses finger vein features extracted from two algorithms namely DWT (Discreet Wavelet Transform) and canny edge detection. Results indicate better accuracy, higher TAR (True acceptance rate), lower TRR (True Reject Rate) and EER (Equal Error Rate) for feature level fusion in comparison with score level fusion technique.

Multimodal Biometric Systems: A Comparative Study

Biometrics technology stands as one of the major backbones that had united biosciences and technology representing an instrument for security and forensics researchers to develop more accurate, robust and confident systems. Starting from uni-modal biometrics as finger print, face, speech and iris passing through multimodal biometrics based on uni-biometrics fused by different fusion techniques as feature level, score level and decision level fusion techniques, biometrics were still one of the most investigated technologies. From here in this paper, we tried to build the base for researchers whom are interested in biometric systems through introducing a comparative study of most used and known uni- and multimodal biometrics such as face, iris, finger vein, face and iris multimodal, face, finger print and finger vein multimodal. Through this comparative study, a comparative model is based on principal component analysis feature extractor and Euclidean distance matcher applied using MATLAB. This model was trained and tested in two different modes homogenous data using SDUMLA-HMT database and heterogeneous mode extracting 106 frontal single face image from CASIA-FACEV5 while the reminder biometrics under consideration from SDUMLA-HMT. Feature level and score level fusions were tested in both modes on all multimodal systems under consideration.

Kernel Locality Preserving Symmetrical Weighted Fisher Discriminant Analysis based subspace approach for expression recognition

This paper mainly focuses on dimensional reduction of fused dataset of holistic and geometrical face features vectors by solving singularity problem of linear discriminant analysis and maximizing the Fisher ratio in nonlinear subspace region with the preservation of local discriminative features. The combinational feature vector space is projected into low dimensional subspace using proposed Kernel Locality Preserving Symmetrical Weighted Fisher Discriminant Analysis (KLSWFDA) method. Matching score level fusion technique has been applied on projected subspace and combinational entire Gabor subspace is framed. Euclidean distance metric (L2) and support vector machine (SVM) classifier has been implemented to recognize and classify the expressions. Performance of proposed approach is evaluated and compared with state of art approaches. Experimental results on JAFFE, YALE and FD expression database demonstrate the effectiveness of the proposed approach.

A robust biometrics multilevel fusion scheme

In this paper, the use of multimodal biometrics is suggested over unimodal biometric system due to the inherent disadvantages of unimodal system. Two or more biometric modalities when combined, offer more reliable identification and verification system. Here, a novel multilevel fusion of palm print and dorsal hand vein is proposed. To fuse the biometric modalities, various levels of fusion are reviewed. In the proposed multilevel fusion, left-and right-hand palm prints of IIT Delhi palmprint database are fused using feature-level fusion to get feature fused vector (FFV). In this, novel feature-level fusion rules are suggested to spatially combine the information of the data samples. These rules are also useful to control the dimension of fused features. After feature-level fusion, features from dorsal hand vein from Bosphorus hand vein database are combined with FFV using score-level fusion. Hence both the feature-level as well as score-level fusion techniques have been used one after the other. The improvement of results verifies the success of our approach. To validate the robustness of proposed multilevel fusion, noise at different level of intensity is artificially added to the test samples. The results suggested the strength of multilevel fusion scheme.

Feature selection for the fusion of face and palmprint biometrics

Multimodal biometric systems aim to improve the recognition accuracy by minimizing the limitations of unimodal systems. In this paper, different fusion schemes based on feature-level and match score-level fusion are employed to provide a robust recognition system. The proposed method presents a multimodal approach based on face–palmprint biometric systems by match score-level fusion technique. Local binary patterns are performed as local feature extractor to obtain efficient texture descriptor. Feature selection is performed using backtracking search algorithm to select an optimal subset of face and palmprint extracted features. Hence, computation time and feature dimension are considerably reduced while obtaining the higher level of performance. Then, match score-level fusion is performed to show the effectiveness and accuracy of the proposed method. In score-level fusion, face and palmprint scores are normalized using tanh normalization and matching scores of individual classifiers are fused using sum rule method. The experimental results are tested on a developed virtual multimodal database combining FERET face and PolyU palmprint databases. The results demonstrate a significant improvement compared with unimodal identifiers, and the proposed approach significantly outperforms other face–palmprint multimodal systems with a recognition accuracy of 99.17 %. Additionally, the proposed approach is compared with the state-of-the-art methods.

Multimodal biometrics: Weighted score level fusion based on non-ideal iris and face images

The iris and face are among the most promising biometric traits that can accurately identify a person because their unique textures can be swiftly extracted during the recognition process. However, unimodal biometrics have limited usage since no single biometric is sufficiently robust and accurate in real-world applications. Iris and face biometric authentication often deals with non-ideal scenarios such as off-angles, reflections, expression changes, variations in posing, or blurred images. These limitations imposed by unimodal biometrics can be overcome by incorporating multimodal biometrics. Therefore, this paper presents a method that combines face and iris biometric traits with the weighted score level fusion technique to flexibly fuse the matching scores from these two modalities based on their weight availability. The dataset use for the experiment is self established dataset named Universiti Teknologi Malaysia Iris and Face Multimodal Datasets (UTMIFM), UBIRIS version 2.0 (UBIRIS v.2) and ORL face databases. The proposed framework achieve high accuracy, and had a high decidability index which significantly separate the distance between intra and inter distance.

Classification of Infant Cries Using Dynamics of Epoch Features

AbstractIn this article, epoch-based dynamic features such as sequence of epoch interval values and epoch strength values are explored to classify infant cries. Epoch is the instant of significant excitation of the vocal tract system during the production of speech. For voiced speech, the most significant excitation takes place around the instant of glottal closure. The different types of infant cries considered in this work are hunger, pain, and wet diaper. In this work, epoch strength and epoch interval features are used to represent infant cry-specific information from the acoustic signal. In this study, the proposed features such as epoch interval and epoch strength values are determined using zero-frequency filter-based method. Gaussian mixture models (GMMs) are used to classify the above-mentioned cries from the features proposed in this work. GMMs are developed separately for each of the cries using the proposed features. The infant cry database collected under a telemedicine project at the Indian Institute of Technology Kharagpur has been used for this study. In the first step, infant cry recognition accuracy is investigated separately using epoch interval and epoch strength features. To enhance recognition performance, GMMs developed using various features are combined through score level fusion techniques. The recognition performance using a combination of evidence is found to be superior over individual systems.

I-vector based speaker recognition using advanced channel compensation techniques

This paper investigates advanced channel compensation techniques for the purpose of improving i-vector speaker verification performance in the presence of high intersession variability using the NIST 2008 and 2010 SRE corpora. The performance of four channel compensation techniques: (a) weighted maximum margin criterion (WMMC), (b) source-normalized WMMC (SN-WMMC), (c) weighted linear discriminant analysis (WLDA) and (d) source-normalized WLDA (SN-WLDA) have been investigated. We show that, by extracting the discriminatory information between pairs of speakers as well as capturing the source variation information in the development i-vector space, the SN-WLDA based cosine similarity scoring (CSS) i-vector system is shown to provide over 20% improvement in EER for NIST 2008 interview and microphone verification and over 10% improvement in EER for NIST 2008 telephone verification, when compared to SN-LDA based CSS i-vector system. Further, score-level fusion techniques are analyzed to combine the best channel compensation approaches, to provide over 8% improvement in DCF over the best single approach, SN-WLDA, for NIST 2008 interview/telephone enrolment-verification condition. Finally, we demonstrate that the improvements found in the context of CSS also generalize to state-of-the-art GPLDA with up to 14% relative improvement in EER for NIST SRE 2010 interview and microphone verification and over 7% relative improvement in EER for NIST SRE 2010 telephone verification.

On the use of genuine-impostor statistical information for score fusion in multimodal biometrics/sur l’usage de l’information statistioue client-imposteur pour la fusion des scores en biométrie multimodale

Matching score level fusion techniques in multimodal person verification conventionally use global score statistics in the normalization and fusion stages. In this paper, novel normalization and fusion methods are presented to take advantage of the separate statistics of the monomodal scores in order to reduce the genuine and impostor pdf lobe overlapping and improve the verification rate. Joint mean normalization is an affine transformation that normalizes the mean of the monomodal biometrics scores separately for the genuine and impostor individuals. Histogram equalization is used to align the statistical distribution of the monomodal scores and make the whole separate statistics comparable. The presented weighting fusion methods have been designed to minimize the variances of the separate multimodal statistics and reduce the multimodal pdf lobe overlapping. The results obtained in speech and face scores fusion upon polycost and xm2vts databases show that the proposed techniques provide better results than the conventional methods.