This code generates a morphing effect between two faces.
- Facial landmarks recognition in both faces using Dlib.
- Triangular Delaunay segmentation.
- Affine transformation between the Delaunay triangles of both faces.
- Alpha blending on the paired triangles.
Steps 3 and 4 are iterated for different values of alpha to generate a bunch of morphing frames.
- After that, frames are converted into a video file using FFmpeg.
- Omit the neck, ear, shoulder points to make it completely automatic.
- Add the option to align and crop the images before morphing.
| Original | Modified |
|---|---|
 |
 |
- Find the max sharpness face in the image with multiple faces.
| Original | Find the max sharpness face |
|---|---|
 |
 |
- Cause this project is running originally on MacOS, the following script installs the dependencies for MacOS.
$ ./install_morphing_dependencies_macos.sh`You can run the script morphing.sh to generate a video file with the morphing effect between two images.
$ ./run_morphing_with_images.sh <image1> <image2> <framerate> <duration_milis> [align]image1: initial image.image2: final image.framerate: frame-rate in fps.duration: morphing duration in miliseconds.align[optional]: if set 1, the images will be aligned before morphing. Default is 0.
I suppose you to open face alignment and cropping option for images with multiple faces to make it more robust and fluent. More details about parameters can be found in the script
morphing.shor source code.
- Unaligned version:
$ ./morphing.sh example/harry.jpg example/hermione.jpg 40 3000- Aligned version:
$ ./morphing.sh example/harry.jpg example/hermione.jpg 40 3000 1| unaligned | aligned |
|---|---|
|
 |
You can check the Delaunay and Voronoi diagrams generated for the example images by running the code draw_delaunay.py.
$ python my_code/landmark_detector.py --image <img_path> # generate the landmarks file
$ python my_code/draw_delaunay.py --image <img_path> # generate the Delaunay and Voronoi diagrams