Temporal Convolutional Neural Networks for Radar Micro-Doppler Based Gait Recognition

Abstract

The capability of sensors to identify individuals in a specific scenario is a topic of high relevance for sensitive sectors such as public security. A traditional approach involves cameras; however, camera-based surveillance systems lack discretion and have high computational and storing requirements in order to perform human identification. Moreover, they are strongly influenced by external factors (e.g., light and weather). This paper proposes an approach based on a temporal convolutional deep neural networks classifier applied to radar micro-Doppler signatures in order to identify individuals. Both sensor and processing requirements ensure a low size weight and power profile, enabling large scale deployment of discrete human identification systems. The proposed approach is assessed on real data concerning 106 individuals. The results show good accuracy of the classifier (the best obtained accuracy is 0.89 with an F1-score of 0.885) and improved performance when compared to other standard approaches.

Monitoring of Neuroendocrine Changes in Acute Stage of Severe Craniocerebral Injury by Transcranial Doppler Ultrasound Image Features Based on Artificial Intelligence Algorithm

“…The results showed that after processing by artificial intelligence neighborhood segmentation algorithm, the accuracy of edge division of injury site in color Doppler ultrasound images of patients with brain injury was significantly improved, enhancing the application value of color Doppler ultrasound images in monitoring patients with severe head injury. However, the operation time of artificial intelligence neighborhood segmentation algorithm in processing transcranial Doppler ultrasound was 3.14 ± 1.02 s, which was significantly shorter than 32.23 ± 9.56 s of the traditional CNN algorithm, with statistical significance ( P < 0.05) indicating that compared with the traditional CNN algorithm, the image processing time of the artificial intelligence neighborhood segmentation algorithm was significantly shortened, greatly improving the work efficiency, which was consistent with the study results of Addabbo et al [ 23 ].…”

Dual-Biometric Human Identification Using Radar Deep Transfer Learning

“…We see that most of the recent results reported in the literature for the radar-based identification are based on the gait biometric [ 4 , 15 , 16 , 17 , 29 , 30 , 31 , 32 , 33 ]. We can relate that to the superior accuracies achieved using the gait signature compared to other signatures, as in Ref.…”

“…Gait is considered a unique signature due to the physical and behavioral characteristics difference between individuals, which can be used as a biometric for human identification. As observed in the literature [ 16 , 17 , 18 ], micro-Doppler-based methods have been implemented using a variety of machine learning structures with the purpose of identifying human targets based on their gait signatures. The main challenges are to maximize the accuracy of identification, minimize the need for larger training datasets and minimize the limitations imposed on the implementation scenarios.…”

Human Gait Recognition: A Single Stream Optimal Deep Learning Features Fusion

“…Through these methods, the human gait can be analyzed, for example, through muscle movement while walking. Gait recognition algorithms have progressed to the point that they can now be used in a wide range of "real-world" applications, such as video monitoring, crime prevention, and forensic detection [13,14].…”

“…Computer vision, a floor sensor, and a watch sensor are three common approaches for detecting gait depending on the background [14]. Using computer vision, cameras are used for video capturing to process the frames' transformation.…”