https://github.com/deepinsight/mxnet-SSH
Reproduce SSH (Single Stage Headless Face Detector) with MXNet
https://github.com/deepinsight/mxnet-SSH
face-detection mxnet
Last synced: 27 days ago
JSON representation
Reproduce SSH (Single Stage Headless Face Detector) with MXNet
- Host: GitHub
- URL: https://github.com/deepinsight/mxnet-SSH
- Owner: deepinsight
- License: mit
- Created: 2018-07-30T15:35:24.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2018-07-31T08:02:37.000Z (over 6 years ago)
- Last Synced: 2024-10-28T07:41:46.382Z (6 months ago)
- Topics: face-detection, mxnet
- Language: Python
- Size: 219 KB
- Stars: 91
- Watchers: 10
- Forks: 23
- Open Issues: 15
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-MXNet - SSH
README
# mxnet-SSH
Reproduce [SSH](https://arxiv.org/abs/1708.03979) (Single Stage Headless Face Detector) with MXNet.
Original Caffe code: [https://github.com/mahyarnajibi/SSH](https://github.com/mahyarnajibi/SSH)
Evaluation on WIDER FACE:
| Impelmentation | Easy-Set | Medium-Set | Hard-Set |
| ------------------ | -------- | ---------- | -------- |
| Original Caffe SSH | 0.93123 | 0.92106 | 0.84582 |
| Our SSH Model | 0.93489 | 0.92281 | 0.84525 |
### Installation
1. Clone the repository.
2. Download MXNet VGG16 pretrained model from [here](http://data.dmlc.ml/models/imagenet/vgg/vgg16-0000.params) and put it under `model` directory.
3. Type `make` to build necessary cxx libs.
### Training
```
python train_ssh.py
```
For more information, check our insightface repository.