Abstract | ||
---|---|---|
Misalignment caused by poorly detected bounding boxes or varying poses of human bodies is a critical challenge to robust person re-identification (re-ID) systems. Most previous works extract features either from the detected entire pedestrian images or from local image patches that are globally aligned or cropped according to human body landmarks. However, these methods still suffer from misalignment ofhuman bodies in different images. In this paper, we present a Deep Joint Learning (DJL) network to fulfill misalignment robust person re-ID. It locally aligns the human bodies by pooling the features around the body parts on feature maps, and jointly optimizes the global and aligned local features to further enhance the discriminative capability oflearned feature representations. Experimental results on Market-1501 and CUHK03 datasets show that our method can effectively handle the misalignment induced intra-class variations and yield competitive accuracy particularly on poorly aligned pedestrian images. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/ISBA.2018.8311470 | 2018 IEEE 4th International Conference on Identity, Security, and Behavior Analysis (ISBA) |
Keywords | Field | DocType |
CUHK03 dataset,Market-1501 dataset,misalignment robust person reidentification,deep joint learning network,pedestrian image detection,bounding box detection,feature representations,pooling body parts,poorly aligned pedestrian images,global aligned local features,feature maps,human body landmarks,local image patches | Pattern recognition,Computer science,Pooling,Robustness (computer science),Feature extraction,Pose,Artificial intelligence,Discriminative model,Bounding overwatch | Conference |
ISBN | Citations | PageRank |
978-1-5386-2249-0 | 0 | 0.34 |
References | Authors | |
13 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yuting Liu | 1 | 8 | 2.15 |
Qijun Zhao | 2 | 419 | 38.37 |
Zhihong Wu | 3 | 16 | 1.01 |