Published on Mon Jul 02 2018

SphereReID: Deep Hypersphere Manifold Embedding for Person Re-Identification

Xing Fan, Wei Jiang, Hao Luo, Mengjuan Fei

Many current successful Person Re-Identification(ReID) methods train a model with the softmax loss function to classify images of different persons. In this paper, we use a modified softmax function to solve the classification problem and learn ahypersphere manifold embedding simultaneously.

0
0
0
Abstract

Many current successful Person Re-Identification(ReID) methods train a model with the softmax loss function to classify images of different persons and obtain the feature vectors at the same time. However, the underlying feature embedding space is ignored. In this paper, we use a modified softmax function, termed Sphere Softmax, to solve the classification problem and learn a hypersphere manifold embedding simultaneously. A balanced sampling strategy is also introduced. Finally, we propose a convolutional neural network called SphereReID adopting Sphere Softmax and training a single model end-to-end with a new warming-up learning rate schedule on four challenging datasets including Market-1501, DukeMTMC-reID, CHHK-03, and CUHK-SYSU. Experimental results demonstrate that this single model outperforms the state-of-the-art methods on all four datasets without fine-tuning or re-ranking. For example, it achieves 94.4% rank-1 accuracy on Market-1501 and 83.9% rank-1 accuracy on DukeMTMC-reID. The code and trained weights of our model will be released.

Wed Dec 25 2019
Computer Vision
Ranking and Classification driven Feature Learning for Person Re_identification
Person re-identification has attracted many researchers' attention for its wide application. Most of current methods uses CNN to learn to embeddings that can capture semantic similarity among data points. This paper will combine the ranking motivated structured loss, Triplet loss and a new metric learning loss function.
0
0
0
Tue Apr 24 2018
Computer Vision
Homocentric Hypersphere Feature Embedding for Person Re-identification
Person re-identification (Person ReID) is a challenging task due to the large variations in camera viewpoint, lighting, resolution, and human pose. Traditional triplet loss methods focus on samples within a mini-batch and lack knowledge of global feature distribution. To address these issues,
0
0
0
Wed Jul 31 2019
Computer Vision
Self-training with progressive augmentation for unsupervised cross-domain person re-identification
Person re-identification (Re-ID) has achieved great improvement with deep learning and a large amount of labelled training data. We propose a new self-training strategy that progressively augments the model capability by adopting conservative and promoting stages.
0
0
0
Fri Jan 29 2021
Computer Vision
Complementary Pseudo Labels For Unsupervised Domain Adaptation On Person Re-identification
Supervised person re-identification (re-ID) models suffer dramatic performance drop when tested on an unseen domain. Existing methods are primary to use pseudo labels to alleviate this problem. We propose a joint learning framework to learn better feature embeddings.
0
0
0
Fri Mar 24 2017
Computer Vision
Scalable Person Re-identification on Supervised Smoothed Manifold
Most existing person re-identification algorithms either extract robust features or learn discriminative metrics for person images. However, the underlying manifold which those images reside on is rarely investigated. An algorithm is proposed that can scale up to large repositories with low on-line time complexity.
0
0
0
Wed Nov 16 2016
Computer Vision
Deep Transfer Learning for Person Re-identification
Person re-identification (Re-ID) poses a unique challenge to deep learning. It is difficult to learn a deep model with millions of parameters on a small training set. A number of deep transfer learning models are proposed to address the data sparsity problem.
0
0
0