Skip to content

Latest commit

 

History

History
67 lines (48 loc) · 1.91 KB

README.md

File metadata and controls

67 lines (48 loc) · 1.91 KB
Raw

2D and 3D Face alignment library build using pytorch

Input

Input

(from https://github.com/1adrianb/face-alignment/tree/master/test/assets)

Ailia input shape : (1, 3, 256, 256) Range : [0.0, 1.0]

Output

  • 2D mode output image
    2D_Output

  • 3D mode output image
    3D_Output

  • confidence map
    confidence_map

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 face_alignment.py

or if you want to try 3D-mode,

$ python3 face_alignment.py --active_3d

will give you back 3D face alignment results.

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.
Confidence map is saved with the output image file name prefixed with _confidence.

  • ex. output.png --> output_confidence.png
$ python3 face_alignment.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

By adding the --video option, you can input the video.
If you pass 0 as an argument to VIDEO_PATH, you can use the webcam input instead of the video file.

$ python3 face_alignment.py --video VIDEO_PATH

Reference

2D and 3D Face alignment library build using pytorch

Framework

Pytorch 1.2.0

Model Format

ONNX opset = 10

Netron