Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
You can not select more than 25 topics
Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
JiahongX
4e438cc743
|
1 year ago | |
|---|---|---|
| config | 1 year ago | |
| configs | 1 year ago | |
| data | 1 year ago | |
| doc | 1 year ago | |
| engine | 1 year ago | |
| imgs | 1 year ago | |
| layers | 1 year ago | |
| logs | 1 year ago | |
| modeling | 1 year ago | |
| scripts | 1 year ago | |
| solver | 1 year ago | |
| tests | 1 year ago | |
| tools | 1 year ago | |
| utils | 1 year ago | |
| .gitattributes | 1 year ago | |
| readme.md | 1 year ago | |
| resnet50-19c8e357.pth | 1 year ago | |
| test.py | 1 year ago | |
readme.md
项目中文名
用于红外可见人员重新识别的特征束自解耦
项目英文名
Feature Bundle Self-disentanglement for Infrared-Visible Person Re-Identification
作者
Diangang Li, Xing Wei, Xiaopeng Hong, Yihong Gong
项目简介
1. 功能
跨模态行人重识别，实现在RGB图像和近红外(infrared)图像之间的检索
2. 性能
在和RegDB上达到当前最优性能
在RegDB上，rank 1 = 74.02%，mAP = 70.51%；
3. 评估指标
rank1
mAP
4. 使用数据集
RegDB
运行环境与依赖
| 类别 | 名称 | 版本 |
|---|---|---|
| os | ubuntu | 16.04.4 |
| 深度学习框架 | pytorch | 1.0.1 |
输入与输出
| 名称 | 说明 |
|---|---|
| 输入 | 红外图像 |
| 输出 | RGB图像 |
运行方式
在terminal下运行以下命令：
#训练方法
python tools/train.py --config_file='configs/all.yml' MODEL.DEVICE_ID "('your device id')" OUTPUT_DIR "('your path to save checkpoints and logs')"
#测试方法
python test.py -data_dir <directory of processed data> --save_dir <directory of log and model>