add readme.md

Author: Liana

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "caffe"]
+	path = caffe
+	url = git://github.com/zlmzju/caffe.git

readme.md

@@ -0,0 +1,32 @@
+# Part-Aligned Network for Person Re-identification
+
+Fangfang Wang*, [Liming Zhao](http://www.zhaoliming.net)*, [Xi Li](http://mypage.zju.edu.cn/xilics), Xinchao Wang, and Dacheng Tao. “[Geometry-Aware Scene Text Detection with Instance Transformation Network](http://openaccess.thecvf.com/content_cvpr_2018/html/Wang_Geometry-Aware_Scene_Text_CVPR_2018_paper.html).” In CVPR, 2018. (*authors contributed equally)
+
+
+
+```
+@InProceedings{Wang_2018_CVPR,
+author = {Wang, Fangfang and Zhao, Liming and Li, Xi and Wang, Xinchao and Tao, Dacheng},
+title = {Geometry-Aware Scene Text Detection with Instance Transformation Network},
+booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
+pages = {1381-1389},
+year = {2018}
+}
+```
+
+Contact: Fangfang Wang ([email protected]),Liming Zhao ([email protected])
+
+## Instructions
+
+- Use my [`Caffe`](https://github.com/zlmzju/caffe/tree/itn) for using transformation layer.
+
+- Run `train.sh` in the `experiments` folder to train the model.
+
+
+## Descriptions
+
+- Use `Caffe` for implementation, please refer to the [Caffe project website](http://caffe.berkeleyvision.org/) for installation.
+
+- The protocal files are located in `experiments/xxx/proto` folder.
+
+- Trained models are located in `experiments/xxx/weight` folder.