Horizontal Pyramid Matching for Person Re-Identification

Authors: Yang Fu, Yunchao Wei, Yuqian Zhou, Honghui Shi, Gao Huang, Xinchao Wang, Zhiqiang Yao, Thomas Huang8295-8302

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To validate the effectiveness of our proposed HPM method, extensive experiments are conducted on three popular datasets including Market-1501, Duke MTMCRe ID and CUHK03.
Researcher Affiliation Collaboration Yang Fu,1 Yunchao Wei,1 Yuqian Zhou,1 Honghui Shi,1,2 Gao Huang,3 Xinchao Wang,4 Zhiqiang Yao,5 Thomas Huang1 1IFP, Beckman Institute, UIUC, 2 IBM Research, 3Cornell University, 4Stevens Institute of Technology, 5Cloud Walk Technology
Pseudocode No The paper describes its method through text and diagrams (Figure 1, Figure 3), but it does not include any pseudocode or algorithm blocks.
Open Source Code No The paper does not include any explicit statement about open-sourcing its code or provide a link to a code repository.
Open Datasets Yes Market1501 (Zheng et al. 2015) contains 32,668 images of 1,501 labeled persons... Duke MTMC-Re ID (Ristani et al. 2016; Zheng, Zheng, and Yang 2017c) is a subset of the Duke MTMC dataset... CUHK03 (Li et al. 2014) consists of 14,097 cropped images from 1,467 identities.
Dataset Splits Yes Market-1501 ... There are 19,732 gallery images and 12,936 training images detected by DPM... including 751 identities in the training set and 750 identities in the testing set. Duke MTMC-Re ID ... There are 2,228 query images, 16,522 training images and 17,661 gallery images... CUHK03 ... adopt the new training/testing protocol proposed in (Zhong et al. 2017).
Hardware Specification Yes Our model is implemented on Pytorch platform and train with two NVIDIA TITAN X GPUs.
Software Dependencies No The paper mentions "Pytorch platform" and "Resnet50 that initialized with the weights pretrained on Image Net". However, it does not specify version numbers for PyTorch or other software libraries/dependencies.
Experiment Setup Yes we resize all the image to 384x128. ... The batch size is set to 64 and we train model for 60 epoch. The base learning rate is set to 0.1 and decay to 0.01 after 40 epochs. ... The stochastic gradient descent (SGD) with 0.9 momentum is implemented in each mini-batch...