Open AccessPerson re‐identification based on pose angle estimation and multi‐feature extraction
Author(s) -
Andavar Suruliandi,
Elango Poongothai
Publication year - 2018
Publication title -
iet biometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.434
H-Index - 28
ISSN - 2047-4946
DOI - 10.1049/iet-bmt.2016.0198
Subject(s) - computer science , artificial intelligence , identification (biology) , matching (statistics) , benchmark (surveying) , pattern recognition (psychology) , process (computing) , pose , computer vision , disjoint sets , feature extraction , feature (linguistics) , feature matching , mathematics , statistics , linguistics , botany , philosophy , geodesy , combinatorics , biology , geography , operating system
Re‐identification enables the tracking of the person taken from different disjoint camera aspects either from online or retrospectively for recognition of his or her visual appearance. Here a new method is proposed for person re‐identification, taking into consideration the pose of the person as the primary factor, with multiple features being extracted from significant portions. Then angle‐based pose priority has applied for matching and identification more robust to viewpoint. Their proposed method helps to reduce the number of images which are redundant in the training phase as well as the number of matching process in the test phase. The strength of the proposed method is demonstrated on three different benchmark databases containing more than 1000 person‐images under variations in illumination, viewpoint and occlusion. The experimental results show that the proposed approach provides a higher recognition rates for all the issues of identification process. Finally, the results prove the superiority of the proposed method over other re‐identification methods both in terms of visual and quantitative comparisons.