In the recent past, deep learning methods have demonstrated remarkable success for supervised learning tasks in multiple domains including computer vision, natural language processing, and speech processing. In this article, we investigate the impact of deep learning in the field of biometrics, given its success in other domains. Since biometrics deals with identifying people by using their characteristics, it primarily involves supervised learning and can leverage the success of deep learning in other related domains. In this article, we survey 100 different approaches that explore deep learning for recognizing individuals using various biometric modalities. We find that most deep learning research in biometrics has been focused on face and speaker recognition. Based on inferences from these approaches, we discuss how deep learning methods can benefit the field of biometrics and the potential gaps that deep learning approaches need to address for real-world biometric applications.

Deep Learning for Biometrics: A Survey

Compared to other biometrics, gait is difficult to conceal and has the advantage of being unobtrusive. Inertial sensors, such as accelerometers and gyroscopes, are often used to capture gait dynamics. These inertial sensors are commonly integrated into smartphones and are widely used by the average person, which makes gait data convenient and inexpensive to collect. In this paper, we study gait recognition using smartphones in the wild. In contrast to traditional methods, which often require a person to walk along a specified road and/or at a normal walking speed, the proposed method collects inertial gait data under unconstrained conditions without knowing when, where, and how the user walks. To obtain good person identification and authentication performance, deep-learning techniques are presented to learn and model the gait biometrics based on walking data. Specifically, a hybrid deep neural network is proposed for robust gait feature representation, where features in the space and time domains are successively abstracted by a convolutional neural network and a recurrent neural network. In the experiments, two datasets collected by smartphones for a total of 118 subjects are used for evaluations. The experiments show that the proposed method achieves higher than 93.5% and 93.7% accuracies in person identification and authentication, respectively.

Deep Learning-Based Gait Recognition Using Smartphones in the Wild

Recognizing people by their gait has become more and more popular nowadays due to the following reasons. First, gait recognition can work well remotely. Second, gait recognition can be done from low-resolution videos and with simple instrumentation. Third, gait recognition can be done without the cooperation of individuals. Fourth, gait recognition can work well while other features such as faces and fingerprints are hidden. Finally, gait features are typically difficult to be impersonated. Recent ubiquity of smartphones that capture gait patterns through accelerometers and gyroscope and advances in machine learning have opened new research directions and applications in gait recognition. A timely survey that addresses current advances is missing. In this article, we survey research works in gait recognition. In addition to recognition based on video, we address new modalities, such as recognition based on floor sensors, radars, and accelerometers; new approaches that include machine learning methods; and examine challenges and vulnerabilities in this field. In addition, we propose a set of future research directions. Our review reveals the current state-of-art and can be helpful to both experts and newcomers of gait recognition. Moreover, it lists future works and publicly available databases in gait recognition for researchers.

A Survey on Gait Recognition

Improved gait recognition based on specialized deep convolutional neural network

Biometric systems are used for the verification and identification of individuals using their physiological or behavioral features. These features can be categorized into unimodal and multimodal systems, in which the former have several deficiencies that reduce the accuracy of the system, such as noisy data, inter-class similarity, intra-class variation, spoofing, and non-universality. However, multimodal biometric sensing and processing systems, which make use of the detection and processing of two or more behavioral or physiological traits, have proved to improve the success rate of identification and verification significantly. This paper provides a detailed survey of the various unimodal and multimodal biometric sensing types providing their strengths and weaknesses. It discusses the stages involved in the biometric system recognition process and further discusses multimodal systems in terms of their architecture, mode of operation, and algorithms used to develop the systems. It also touches on levels and methods of fusion involved in biometric systems and gives researchers in this area a better understanding of multimodal biometric sensing and processing systems and research trends in this area. It furthermore gives room for research on how to find solutions to issues on various unimodal biometric systems.

Unimodal and Multimodal Biometric Sensing Systems: A Review

Gait recognition is a biometric technique that is used in order to determine the identity of humans based on the style and the manner of their walk. Yet, Gait recognition performance is often degraded by some covariate factors such as a viewing angle variation, clothing and carrying condition changes, and a low-image resolution. In general, current tactics to object recognition highly depend on the use of machine learning techniques. Therefore, a deep convolutional neural network (CNN) is one of the most advanced machine learning techniques that has the ability to approximate complex non-linear functions from high-dimensional input data in a hierarchical process. In this paper, we develop a specialized deep CNN architecture, which consists of multilayers of convolutional and subsampling layers. The proposed technique is less sensitive to several cases of the common variations and occlusions that affect and degrade gait recognition performance. We avoided the use of the typical subspace learning methods, along with its shortcomings, that are widely used in gait recognition. When applied the proposed deep CNN model to CASIA-B large gait database, the experimental results show that the deep CNN model developed in this paper outperforms the other state of art gait recognition techniques in several cases.

Improved Gait recognition based on specialized deep convolutional neural networks

One of the most challenging topics in the field of intelligent transportation systems is the automatic interpretation of the driver’s behavior. This research investigates distracted driver posture recognition as a part of the human action recognition framework. Numerous car accidents have been reported that were caused by distracted drivers. Our aim was to improve the performance of detecting drivers’ distracted actions. The developed system involves a dashboard camera capable of detecting distracted drivers through 2D camera images. We use a combination of three of the most advanced techniques in deep learning, namely the inception module with a residual block and a hierarchical recurrent neural network to enhance the performance of detecting the distracted behaviors of drivers. The proposed method yields very good results. The distracted driver behaviors include texting, talking on the phone, operating the radio, drinking, reaching behind, fixing hair and makeup, and talking to the passenger.

Distracted driver classification using deep learning

Over the past few decades, hyperspectral image (HSI) classification has garnered increasing attention from the remote sensing research community. The largest challenge faced by HSI classification is the high feature dimensions represented by the different HSI bands given the limited number of labeled samples. Deep learning and convolutional neural networks (CNNs), in particular, have been shown to be highly effective in several computer vision problems such as object detection and image classification. In terms of accuracy and computational cost, one of the best CNN architectures is the Inception model i.e., the winner of the ImageNet Large Scale Visual Recognition Competition (ILSVRC) 2014 challenge. Another architecture that has significantly improved image recognition performance is the Residual Network (ResNet) architecture i.e., the winner of the ILSVRC 2015 challenge. Inspired by the incredible performance introduced by the Inception and ResNet architectures, we investigate the possibility of combining the core ideas of these two models into a hybrid architecture to improve the HSI classification performance. We tested this combined model on four standard HSI datasets, and it shows competitive results compared with other existing HSI classification methods. Our hybrid deep ResNet-Inception architecture obtained accuracies of 95.31% on the Pavia University dataset, 99.02% on the Pavia Centre scenes dataset, 95.33% on the Salinas dataset and 90.57% on the Indian Pines dataset.

A Hybrid Deep ResNet and Inception Model for Hyperspectral Image Classification

Internet of things (IoT) devices and applications are dramatically increasing worldwide, resulting in more cybersecurity challenges. Among these challenges are malicious activities that target IoT devices and cause serious damage, such as data leakage, phishing and spamming campaigns, distributed denial-of-service (DDoS) attacks, and security breaches. In this paper, a stacked deep learning method is proposed to detect malicious traffic data, particularly malicious attacks targeting IoT devices. The proposed stacked deep learning method is bundled with five pretrained residual networks (ResNets) to deeply learn the characteristics of the suspicious activities and distinguish them from normal traffic. Each pretrained ResNet model consists of 10 residual blocks. We used two large datasets to evaluate the performance of our detection method. We investigated two heterogeneous IoT environments to make our approach deployable in any IoT setting. Our proposed method has the ability to distinguish between benign and malicious traffic data and detect most IoT attacks. The experimental results show that our proposed stacked deep learning method can provide a higher detection rate in real time compared with existing classification techniques.

/pdf/a-stacked-deep-learning-approach-for-iot-cyberattack-1qwcxee9w3.pdf

Munif Alotaibi

Papers

Improved gait recognition based on specialized deep convolutional neural network

Improved Gait recognition based on specialized deep convolutional neural networks

Distracted driver classification using deep learning

A Hybrid Deep ResNet and Inception Model for Hyperspectral Image Classification

A Stacked Deep Learning Approach for IoT Cyberattack Detection