https://github.com/akofman/wider-face-pascal-voc-annotations
WIDER FACE annotations converted to the Pascal VOC XML format
https://github.com/akofman/wider-face-pascal-voc-annotations
dataset face-detection machine-learning pascal-voc widerface
Last synced: about 1 year ago
JSON representation
WIDER FACE annotations converted to the Pascal VOC XML format
- Host: GitHub
- URL: https://github.com/akofman/wider-face-pascal-voc-annotations
- Owner: akofman
- License: mit
- Created: 2018-07-23T21:33:16.000Z (almost 8 years ago)
- Default Branch: master
- Last Pushed: 2023-10-03T21:12:35.000Z (over 2 years ago)
- Last Synced: 2025-04-23T03:48:27.937Z (about 1 year ago)
- Topics: dataset, face-detection, machine-learning, pascal-voc, widerface
- Language: Python
- Homepage:
- Size: 5.93 MB
- Stars: 80
- Watchers: 2
- Forks: 49
- Open Issues: 7
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
README
# WIDER FACE PASCAL VOC ANNOTATIONS
This repository contains the [WIDER FACE](http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/) annotations converted to the [Pascal VOC](http://host.robots.ox.ac.uk/pascal/VOC/) XML format.
It also contains the script used to do this conversion.
```
usage: convert.py [-h] [-ap ANNOTATIONS_PATH] [-tp TARGET_PATH]
[-ip IMAGES_PATH]
optional arguments:
-h, --help show this help message and exit
-ap ANNOTATIONS_PATH, --annotations-path ANNOTATIONS_PATH
the annotations file path.
ie:"./wider_face_split/wider_face_train_bbx_gt.txt".
-tp TARGET_PATH, --target-path TARGET_PATH
the target directory path where XML files will be
copied.
-ip IMAGES_PATH, --images-path IMAGES_PATH
the images directory path. ie:"./WIDER_train/images"
```
It has been used as following:
```
$ ./convert.py -ap ./wider_face_split/wider_face_train_bbx_gt.txt -tp ./WIDER_train_annotations/ -ip ./WIDER_train/images/
$ ./convert.py -ap ./wider_face_split/wider_face_val_bbx_gt.txt -tp ./WIDER_val_annotations/ -ip ./WIDER_val/images/
```
# License
MIT Licensed. Copyright (c) Alexis Kofman 2018.