Efficient Identification in Large-Scale Vein Recognition Systems Using Spectral Minutiae Representations

Speaker recognition deals with recognizing speakers by their speech. Most speaker recognition systems are built upon two stages, the first stage extracts low dimensional correlation embeddings from speech, and the second performs the classification task. The robustness of a speaker recognition system mainly depends on the extraction process of speech embeddings, which are primarily pre-trained on a large-scale dataset. As the embedding systems are pre-trained, the performance of speaker recognition models greatly depends on domain adaptation policy, which may reduce if trained using inadequate data. This paper introduces a speaker recognition strategy dealing with unlabeled data, which generates clusterable embedding vectors from small fixed-size speech frames. The unsupervised training strategy involves an assumption that a small speech segment should include a single speaker. Depending on such a belief, a pairwise constraint is constructed with noise augmentation policies, used to train AutoEmbedder architecture that generates speaker embeddings. Without relying on domain adaption policy, the process unsupervisely produces clusterable speaker embeddings, termed unsupervised vectors (u-vectors). The evaluation is concluded in two popular speaker recognition datasets for English language, TIMIT, and LibriSpeech. Also, a Bengali dataset is included to illustrate the diversity of the domain shifts for speaker recognition systems. Finally, we conclude that the proposed approach achieves satisfactory performance using pairwise architectures.

Download Full-text

Local feature extraction based facial emotion recognition: a survey

International Journal of Electrical and Computer Engineering (IJECE) ◽

10.11591/ijece.v10i4.pp4080-4092 ◽

2020 ◽

Vol 10 (4) ◽

pp. 4080

Author(s):

Khadija Slimani ◽

Mohamed Kas ◽

Youssef El Merabet ◽

Yassine Ruichek ◽

Rochdi Messoussi

Keyword(s):

Facial Expression ◽

Large Scale ◽

State Of The Art ◽

Research Work ◽

Emotional Expressions ◽

Expression Recognition ◽

Technological Advancement ◽

Recognition Systems ◽

Local Feature Extraction ◽

Micro Patterns

Notwithstanding the recent technological advancement, the identification of facial and emotional expressions is still one of the greatest challenges scientists have ever faced. Generally, the human face is identified as a composition made up of textures arranged in micro-patterns. Currently, there has been a tremendous increase in the use of local binary pattern based texture algorithms which have invariably been identified to being essential in the completion of a variety of tasks and in the extraction of essential attributes from an image. Over the years, lots of LBP variants have been literally reviewed. However, what is left is a thorough and comprehensive analysis of their independent performance. This research work aims at filling this gap by performing a large-scale performance evaluation of 46 recent state-of-the-art LBP variants for facial expression recognition. Extensive experimental results on the well-known challenging and benchmark KDEF, JAFFE, CK and MUG databases taken under different facial expression conditions, indicate that a number of evaluated state-of-the-art LBP-like methods achieve promising results, which are better or competitive than several recent state-of-the-art facial recognition systems. Recognition rates of 100%, 98.57%, 95.92% and 100% have been reached for CK, JAFFE, KDEF and MUG databases, respectively.

Download Full-text

Robustness Evaluation of Hand Vein Recognition Systems

2015 International Conference of the Biometrics Special Interest Group (BIOSIG) ◽

10.1109/biosig.2015.7314619 ◽

2015 ◽

Cited By ~ 7

Author(s):

Christof Kauba ◽

Andreas Uhl

Keyword(s):

Hand Vein ◽

Recognition Systems ◽

Vein Recognition

Download Full-text

Methods for accuracy‐preserving acceleration of large‐scale comparisons in CPU‐based iris recognition systems

IET Biometrics ◽

10.1049/iet-bmt.2016.0125 ◽

2017 ◽

Vol 7 (4) ◽

pp. 356-364 ◽

Cited By ~ 1

Author(s):

Christian Rathgeb ◽

Nicolas Buchmann ◽

Heinz Hofbauer ◽

Harald Baier ◽

Andreas Uhl ◽

...

Keyword(s):

Iris Recognition ◽

Large Scale ◽

Recognition Systems

Download Full-text

Finger Vein Recognition System Based on Convolutional Neural Network and Android

Journal of Physics Conference Series ◽

10.1088/1742-6596/2078/1/012053 ◽

2021 ◽

Vol 2078 (1) ◽

pp. 012053

Author(s):

Yangfeng Wang ◽

Tao Chen

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Large Scale ◽

Rapid Development ◽

Real Life ◽

Recognition System ◽

Finger Vein ◽

Recognition Function ◽

Vein Recognition ◽

Finger Vein Recognition

Abstract With the rapid development of science and technology, biotechnology has developed rapidly. Among the many biometric technologies, finger vein technology has the characteristics of vitality, portability, and non-replicability, so it is considered to be the most promising biometric technology. However, the accuracy of finger vein recognition is affected by the collection device, the surrounding temperature and the algorithm. The flaws cannot be applied to real life on a large scale. This paper designs a finger vein recognition system based on convolutional neural network and Android, which mainly includes the following three parts. First, the system hardware includes the design of the acquisition device, the selection of the core development board and the display screen. Second, the design of the entire system software architecture is based on the MVVM architecture, which ensures low coupling of the program and is easy for later expansion and maintenance. The software includes collection function, recognition function and administrator function. Finally, a lightweight neural network is proposed for finger vein feature extraction, and proposed a storage method based on MMKV to meet the real-time performance of the system.

Download Full-text

Biometric Spoofing and Anti-Spoofing

Developing Next-Generation Countermeasures for Homeland Security Threat Prevention - Advances in Information Security, Privacy, and Ethics ◽

10.4018/978-1-5225-0703-1.ch007 ◽

2017 ◽

pp. 121-139

Author(s):

Zahid Akhtar

Keyword(s):

Large Scale ◽

Security Requirements ◽

Security Systems ◽

Comprehensive Overview ◽

Research Directions ◽

Person Recognition ◽

Biometric Systems ◽

Prime Objective ◽

Recognition Systems ◽

Multimodal Biometric Systems

The demand for reliable and robust person recognition systems has expanded due to intense security requirements in today's highly intertwined network society. The advantages of biometrics over traditional security systems have triggered large-scale deployment of biometrics as an authentic technique to determine the identity of an individual. The prime objective of such methods is to assure that the systems are only accessed by genuine users. Since, biometric traits are overt, leading thus to a threat of them being captured, copied, and forged. Numerous techniques have been developed over the years for biometric spoofing and anti-spoofing. The goal of this chapter is to provide a comprehensive overview on works in the field of spoofing and anti-spoofing with special attention to three mainly accepted biometric traits (i.e., fingerprint, face and iris) and multimodal biometric systems. We also present the key challenges, major issues and point out some of the salient and useful research directions.

Download Full-text

A Selection Module for Large-Scale Face Recognition Systems

Image Analysis and Processing — ICIAP 2015 - Lecture Notes in Computer Science ◽

10.1007/978-3-319-23234-8_49 ◽

2015 ◽

pp. 529-539

Author(s):

Giuliano Grossi ◽

Raffaella Lanzarotti ◽

Jianyi Lin

Keyword(s):

Face Recognition ◽

Large Scale ◽

Recognition Systems

Download Full-text

Learning Meta Model for Zero- and Few-Shot Face Anti-Spoofing

Proceedings of the AAAI Conference on Artificial Intelligence ◽

10.1609/aaai.v34i07.6866 ◽

2020 ◽

Vol 34 (07) ◽

pp. 11916-11923 ◽

Cited By ~ 4

Author(s):

Yunxiao Qin ◽

Chenxu Zhao ◽

Xiangyu Zhu ◽

Zezheng Wang ◽

Zitong Yu ◽

...

Keyword(s):

Face Recognition ◽

Supervised Learning ◽

Large Scale ◽

Training Data ◽

Learning Problem ◽

Meta Model ◽

Meta Learning ◽

Recognition Systems

Face anti-spoofing is crucial to the security of face recognition systems. Most previous methods formulate face anti-spoofing as a supervised learning problem to detect various predefined presentation attacks, which need large scale training data to cover as many attacks as possible. However, the trained model is easy to overfit several common attacks and is still vulnerable to unseen attacks. To overcome this challenge, the detector should: 1) learn discriminative features that can generalize to unseen spoofing types from predefined presentation attacks; 2) quickly adapt to new spoofing types by learning from both the predefined attacks and a few examples of the new spoofing types. Therefore, we define face anti-spoofing as a zero- and few-shot learning problem. In this paper, we propose a novel Adaptive Inner-update Meta Face Anti-Spoofing (AIM-FAS) method to tackle this problem through meta-learning. Specifically, AIM-FAS trains a meta-learner focusing on the task of detecting unseen spoofing types by learning from predefined living and spoofing faces and a few examples of new attacks. To assess the proposed approach, we propose several benchmarks for zero- and few-shot FAS. Experiments show its superior performances on the presented benchmarks to existing methods in existing zero-shot FAS protocols.

Download Full-text

An Improved AM–FM-Based Approach for Reconstructing Fingerprints from Minutiae

International Journal of Image and Graphics ◽

10.1142/s0219467815500072 ◽

2015 ◽

Vol 15 (01) ◽

pp. 1550007

Author(s):

Shuiwang Li ◽

Qijun Zhao ◽

Xiangdong Fei

Keyword(s):

Performance Evaluation ◽

Amplitude Modulation ◽

Large Scale ◽

State Of The Art ◽

Modulation Frequency ◽

Fingerprint Recognition ◽

Reconstruction Methods ◽

Orientation Fields ◽

Recognition Systems ◽

Branch Cuts

Reconstructing fingerprint images from a given set of minutiae is an important issue in analyzing the masquerade attack of automated fingerprint recognition systems (AFRSs) and in generating large scale databases of synthetic fingerprint images for the performance evaluation of AFRSs. Existing fingerprint reconstruction methods either cannot generate visually plausible or realistic fingerprint images, or suffer from the occurrence of false minutiae in the reconstructed fingerprint images. In this paper, we analyze the underlying reason of false minutiae generated by state-of-the-art amplitude modulation–frequency modulation (AM–FM)-based methods. Furthermore, we propose an improved approach by devising a better way to cope with the branch cuts (or discontinuities) in the fingerprint ridge orientation fields, and by introducing an effective scheme to remove false minutiae from the reconstructed fingerprint images. Compared with previous AM–FM based methods, the proposed method gets rid of block effects and successfully reduces the number of false minutiae. Theoretic proofs are provided with respect to the effectiveness of the proposed method for fingerprints with multiple singular points. The proposed method has also been evaluated on public fingerprint databases. The results demonstrate that it is superior to the existing methods in reconstructing realistic fingerprint images with fewer false minutiae.

Download Full-text