Merge pull request #12 from Media-Smart/widerface-dev

update data preparation
Media-Smart · Dec 2, 2020 · a9926db · a9926db
2 parents 93a27c1 + b787212
commit a9926db
Show file tree

Hide file tree

Showing 2 changed files with 36 additions and 7 deletions.
diff --git a/configs/trainval/tinaface/README.md b/configs/trainval/tinaface/README.md
@@ -5,25 +5,51 @@ Providing a strong but simple baseline method for face detection named TinaFace.
 
 ![1](./imgs/model.png)
 
-## Data
+## Data Preparation
 
 ### WIDER FACE
-a. Convert raw data to PASCAL-VOC format by [this tool](https://github.com/akofman/wider-face-pascal-voc-annotations)
+a. Convert raw data to PASCAL-VOC format by this [tool](https://github.com/akofman/wider-face-pascal-voc-annotations)
+
+b. Get *Annotations* folder generated by last step and change the folder structure as follow
+```plain
+vedadet
+├── vedadet
+├── vedacore
+├── tools
+├── configs
+├── data
+│   ├── WIDERFace
+│   │   ├── WIDER_train
+│   │   │   ├── 0--Parade
+│   │   │   ├── ......
+│   │   │   ├── 61--Street_Battle
+│   │   │   ├── Annotations
+│   │   ├── WIDER_val
+│   │   │   ├── 0--Parade
+│   │   │   ├── ......
+│   │   │   ├── 61--Street_Battle
+│   │   │   ├── Annotations
+```
+where *0--Parade*, *61--Street_Battle* and etc. folders contain images.
+
+c. Run the following code to generate txt file of xmls.
+```shell
+cd ${vedadet_root}
+bash configs/trainval/tinaface/gen_xml_name_txt.sh
+```
 
 ## Train
-a. Follow the official instructions on [vedadet](https://github.com/Media-Smart/vedadet) 
+Follow the official instructions on [vedadet](https://github.com/Media-Smart/vedadet) 
 
 ## Evaluation
-a. Back to `${vedadet_root}`
-
-b. Run following instruction
 ```shell
+cd ${vedadet_root}
 python config/trainval/tinaface/test_widerface.py configs/trainval/tinaface/tinaface.py weight_path
 ```
 widerface txt file will be generated at `${vedadet_root}/eval_dirs/tmp/tinaface/`, and then download the [eval_tool](http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/support/eval_script/eval_tools.zip) to evaluate the WIDERFACE performance.
 
 ## Infer
-a. Follow the official instructions on [vedadet](https://github.com/Media-Smart/vedadet) 
+Follow the official instructions on [vedadet](https://github.com/Media-Smart/vedadet) 
 
 ## Results and Models
 

diff --git a/configs/trainval/tinaface/gen_xml_name_txt.sh b/configs/trainval/tinaface/gen_xml_name_txt.sh
@@ -0,0 +1,3 @@
+
+ls -l data/WIDERFace/WIDER_train/Annotations | grep "^-" | awk '{print $9}' | cut -d '.' -f 1 > data/WIDERFace/WIDER_train/train.txt
+ls -l data/WIDERFace/WIDER_val/Annotations | grep "^-" | awk '{print $9}' | cut -d '.' -f 1 > data/WIDERFace/WIDER_val/val.txt
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,3 @@

		ls -l data/WIDERFace/WIDER_train/Annotations \| grep "^-" \| awk '{print $9}' \| cut -d '.' -f 1 > data/WIDERFace/WIDER_train/train.txt
		ls -l data/WIDERFace/WIDER_val/Annotations \| grep "^-" \| awk '{print $9}' \| cut -d '.' -f 1 > data/WIDERFace/WIDER_val/val.txt