kumoh national institute of technology
Networked Systems Lab.

Ali Aouto, Thien Huynh-The, Jae-Min Lee and Dong-Seong Kim, "Pose-Based Identification Using Deep Learning for Military Surveillance Systems", The 10th International Conference on ICT Convergence October 2019 / Ramada Plaza Hotel, Jeju Island, Korea (N8 & N11) (A)
By :
Date : 2019-09-04
Views : 420

Surveillance systems research area these days has become more approachable duo to the increment of crimes and terrorism events, and those systems ideal goal is to be able to identify any person anywhere at anytime, to prevent the crimes before they happen, this paper proposes a gait recognition scheme based on spatiotemporal pose by using skeletal data that was collected from RGB-D sensor and describe that data in as 3D geometrical feature with considering both distances and orientations of joints by studying skeleton sequences over a period of time, after extracting the data it get fed to a fine-tuned pre-trained Convolutional Neural Network (CNN), the approach was tested using the benchmark dataset UPCV Gait. the system evaluates at 92.88% accuracy, which is surprisingly good for using a dataset such UPCV Gait with a fine-tuned pre-traind CNN.