Clustering Matters: Sphere Feature for Fully Unsupervised Person Re-identification
Yi Zheng,Yong Zhou,Jiaqi Zhao,Ying Chen,Rui Yao,Bing Liu,Abdulmotaleb El Saddik
DOI: https://doi.org/10.1145/3501404
IF: 4.094
2022-01-01
ACM Transactions on Multimedia Computing Communications and Applications
Abstract:In person re-identification (Re-ID), the data annotation cost of supervised learning, is huge and it cannot adapt well to complex situations. Therefore, compared with supervised deep learning methods, unsupervised methods are more in line with actual needs. In unsupervised learning, a key to solving Re-ID is to find a standard that can effectively distinguish the difference (distance) between the features of images belonging to different pedestrian identities. However, there are some differences in the images captured by different cameras (such as brightness, angle, etc.). It is well known that the training of neural networks is mainly based on the distance between features, while in unsupervised learning, especially in unsupervised learning methods based on hierarchical clustering, the distance between features plays a more important role in the clustering phase. We improve the accuracy of a deep learning method based on hierarchical clustering under fully unsupervised conditions, starting from both feature and distance metrics. First, we propose to use spherical features, by normalizing the images in the feature space, to weaken the structural differences (length) between features, while saving the feature differences (direction) between different identities. Then, we use the sum of squared errors (SSE) as a regularization term to balance different cluster states. We evaluate our method on four large-scale Re-ID datasets, and experiments show that our method achieves better results than the state-of-the-art unsupervised methods.