PKU-ICST-MIPL / DMA_TIFS2023

Introduction

This is the source code of our TIFS 2024 paper "DMA: Dual Modality-Aware Alignment for Visible-Infrared Person Re-Identification". Please cite the following paper if you use our code.

Zhenyu Cui, Jiahuan Zhou, and Yuxin Peng, "DMA: Dual Modality-Aware Alignment for Visible-Infrared Person Re-Identification", IEEE Transactions on Information Forensics and Security (TIFS), 2024.

Dependencies

• Python 3.7

• cudatoolkit 11.3

• cudnn 8.4

• PyTorch 1.9.1

Data Preparation

• Download the SYSU-MM01 dataset and the RegDB dataset, and place them to /home/cuizhenyu/Dataset_VIReID/ folders.

Usage

• Start training by executing the following commands.

1. For SYSU-MM01 dataset:

   Train: python train.py --dataset sysu -sche 80 140 -v 1 -maxe 200

   Test: python test.py --dataset sysu --model_path ./save_model/sysu_v1/model_best.t

2. For RegDB dataset: python train.py --dataset regdb -sche 80 140 -v 1 -maxe 200

For any questions, feel free to contact us ([e-mail]).

Welcome to our Laboratory Homepage for more information about our papers, source codes, and datasets.