https://github.com/mtli/bffl
Code for Brute-Force Facial Landmark Analysis With A 140,000-Way Classifier :smiley:
https://github.com/mtli/bffl
ai computer-vision facial-landmarks
Last synced: 5 months ago
JSON representation
Code for Brute-Force Facial Landmark Analysis With A 140,000-Way Classifier :smiley:
- Host: GitHub
- URL: https://github.com/mtli/bffl
- Owner: mtli
- License: mit
- Created: 2018-02-06T02:46:27.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2018-02-11T01:54:39.000Z (over 8 years ago)
- Last Synced: 2025-07-07T11:11:17.215Z (12 months ago)
- Topics: ai, computer-vision, facial-landmarks
- Language: Matlab
- Homepage:
- Size: 3.34 MB
- Stars: 37
- Watchers: 3
- Forks: 9
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Brute-Force Facial Landmark Analysis
[[arXiv]](https://arxiv.org/abs/1802.01777)
## Dependency
* VLFeat
* MatConvNet (tested with commit [d62881db](https://github.com/vlfeat/matconvnet/tree/d62881dbb587e4d5ed6750549b6a6b3f7559c84f))
## Usage
1. Download the [pre-trained model](https://drive.google.com/file/d/1oTRRsYseMnIXWiBR-OODuYm1I3NgdDvQ/view?usp=sharing) and extract to `models/`
2. Run `Test.m`
## Face detection
The detection for the example images are provided. However, to run on new images, a face detector is required. We recommend using [MTCNNv2](https://kpzhang93.github.io/MTCNN_face_detection_alignment/) due to its robustness and stability. Also, our detection refinement module is trained with MTCNNv2 using its default parameters.
The accepted format of the bounding box is [x y width height] (no need to round to integer), different from the output of the `detect_face` function in MTCNNv2. It can be transformed using the following code:
```
bbx(:, 3:4) = bbx(:, 3:4) - bbx(:, 1:2);
```
## Videos
* [Temporal smoothing under complete occlusion (provided detection)](doc/HMM.mp4)
* [Interactive conditional prediction - eye corner](doc/Interactive-Eye.mp4)
* [Interactive conditional prediction - nose tip](doc/Interactive-Nose.mp4)
## Citation
If you use this code for your research, please cite the paper:
```
@article{BFFL2018,
title={Brute-Force Facial Landmark Analysis With A 140,000-Way Classifier},
author={Li, Mengtian and Jeni, Laszlo and Ramanan, Deva},
journal={AAAI},
year={2018}
}
```