[MOT] add fairmot_headtracking (#3730)

* add the config of fairmot_headtracking

configs/mot/fairmot/headtracking21/README_cn.md

+简体中文 | [English](README.md)
+
+# FairMOT (FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Tracking)
+
+## 内容
+- [简介](#简介)
+- [模型库](#模型库)
+- [快速开始](#快速开始)
+- [引用](#引用)
+
+## 简介
+
+[FairMOT](https://arxiv.org/abs/2004.01888)以Anchor Free的CenterNet检测器为基础，克服了Anchor-Based的检测框架中anchor和特征不对齐问题，深浅层特征融合使得检测和ReID任务各自获得所需要的特征，并且使用低维度ReID特征，提出了一种由两个同质分支组成的简单baseline来预测像素级目标得分和ReID特征，实现了两个任务之间的公平性，并获得了更高水平的实时多目标跟踪精度。
+
+<div align="center">
+  <img src="../../../../docs/images/ht_fairmot.gif" width='800'/>
+</div>
+
+## 模型库
+
+### FairMOT在HT-21 Training Set上结果
+
+|    骨干网络      |  输入尺寸 |  MOTA  |  IDF1  |  IDS  |   FP  |   FN   |   FPS   |  下载链接 | 配置文件 |
+| :--------------| :------- | :----: | :----: | :---: | :----: | :---: | :------: | :----: |:----: |
+| DLA-34         | 1088x608 |  67.2 |  70.4  |   9403  |  124840  |  255007  |     -   | [下载链接](https://paddledet.bj.bcebos.com/models/mot/fairmot_dla34_30e_1088x608_headtracking21.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml) |
+
+
+### FairMOT在HT-21 Test Set上结果
+
+|    骨干网络      |  输入尺寸 |  MOTA  |  IDF1  |   IDS  |   FP   |   FN   |    FPS   |  下载链接  | 配置文件 |
+| :--------------| :------- | :----: | :----: | :----: | :----: | :----: |:-------: | :----: | :----: |
+| DLA-34         | 1088x608 |  58.2  |  61.3  |  13166   |  141872  |  197074 |    -     | [下载链接](https://paddledet.bj.bcebos.com/models/mot/fairmot_dla34_30e_1088x608_headtracking21.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml) |
+
+**注意:**
+ FairMOT使用8个GPU进行训练，每个GPU上batch size为6，训练30个epoch。
+
+## 快速开始
+
+### 1. 训练
+
+使用8GPU通过如下命令一键式启动训练
+
+```bash
+python -m paddle.distributed.launch --log_dir=./fairmot_dla34_30e_1088x608_headtracking21/ --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml
+```
+
+
+### 2. 评估
+
+使用单张GPU通过如下命令一键式启动评估
+
+```bash
+# 使用PaddleDetection发布的权重
+CUDA_VISIBLE_DEVICES=0 python tools/eval_mot.py -c configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml -o weights=https://paddledet.bj.bcebos.com/models/mot/fairmot_dla34_30e_1088x608_headtracking21.pdparams
+
+# 使用训练保存的checkpoint
+CUDA_VISIBLE_DEVICES=0 python tools/eval_mot.py -c configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml -o weights=output/fairmot_dla34_30e_1088x608_headtracking21/model_final.pdparams
+```
+
+
+### 3. 预测
+
+使用单个GPU通过如下命令预测一个视频，并保存为视频
+
+```bash
+# 预测一个视频
+CUDA_VISIBLE_DEVICES=0 python tools/infer_mot.py -c configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml -o weights=https://paddledet.bj.bcebos.com/models/mot/fairmot_dla34_30e_1088x608_headtracking21.pdparams --video_file={your video name}.mp4  --save_videos
+```
+**注意:**
+ 请先确保已经安装了[ffmpeg](https://ffmpeg.org/ffmpeg.html), Linux(Ubuntu)平台可以直接用以下命令安装：`apt-get update && apt-get install -y ffmpeg`。
+
+### 4. 导出预测模型
+
+```bash
+CUDA_VISIBLE_DEVICES=0 python tools/export_model.py -c configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml -o weights=https://paddledet.bj.bcebos.com/models/mot/fairmot_dla34_30e_1088x608_headtracking21.pdparams
+```
+
+### 5. 用导出的模型基于Python去预测
+
+```bash
+python deploy/python/mot_jde_infer.py --model_dir=output_inference/fairmot_dla34_30e_1088x608_headtracking21 --video_file={your video name}.mp4 --device=GPU --save_mot_txts
+```
+**注意:**
+ 跟踪模型是对视频进行预测，不支持单张图的预测，默认保存跟踪结果可视化后的视频，可添加`--save_mot_txts`表示保存跟踪结果的txt文件，或`--save_images`表示保存跟踪结果可视化图片。
+
+
+## 引用
+```
+@article{zhang2020fair,
+  title={FairMOT: On the Fairness of Detection and Re-Identification in Multiple Object Tracking},
+  author={Zhang, Yifu and Wang, Chunyu and Wang, Xinggang and Zeng, Wenjun and Liu, Wenyu},
+  journal={arXiv preprint arXiv:2004.01888},
+  year={2020}
+}
+@InProceedings{Sundararaman_2021_CVPR,
+    author    = {Sundararaman, Ramana and De Almeida Braga, Cedric and Marchand, Eric and Pettre, Julien},
+    title     = {Tracking Pedestrian Heads in Dense Crowd},
+    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
+    month     = {June},
+    year      = {2021},
+    pages     = {3865-3875}
+}
+```

configs/mot/fairmot/headtracking21/fairmot_dla34_30e_1088x608_headtracking21.yml

+_BASE_: [
+  '../fairmot_dla34_30e_1088x608.yml'
+]
+
+
+# for MOT training
+TrainDataset:
+  !MOTDataSet
+    dataset_dir: dataset/mot
+    image_lists: ['ht21.train']
+    data_fields: ['image', 'gt_bbox', 'gt_class', 'gt_ide']
+
+# for MOT evaluation
+# If you want to change the MOT evaluation dataset, please modify 'task' and 'data_root'
+EvalMOTDataset:
+  !MOTImageFolder
+    dataset_dir: dataset/mot
+    data_root: HT21/images/test
+    keep_ori_im: False # set True if save visualization images or video, or used in DeepSORT
+
+# for MOT video inference
+TestMOTDataset:
+  !MOTVideoDataset
+    dataset_dir: dataset/mot
+    keep_ori_im: True # set True if save visualization images or video