- Standardize image data reading and writing
- Optimize loading (just one pass required)
- Make all sort groups binnable (to greater or lesser results)
- Add sort by pitch
- Deprecate multiple options
- linting, docs + locales
- alignments.py
- Add typed dicts for imported alignments
- Explicitly check for presence of thumb value in alignments dict
- linting
- detected_face.py
- Typing
- Linting
- Legacy support for pre-aligned face
- Update dependencies to new property names
* feat: add support for rectangle cursor shape for masking
* formatting update
* fix bug with initial rectangular point
Co-authored-by: markus <markus@helmes.com>
* Add a new sorting/grouping feature based on the percentage of black pixels in images
Context:
- Faces are quite often moving and can get near borders of the image.
- Because of that, the face image has the out of bound part replaced by black pixels.
- Some are still good for training but others not. Also, one might want to keep an alignment file with the near borders faces but excluded (some of) them for training.
- Having to manually move/delete those images in a dataset of thousands of them is painful and prone to mistakes.
- Thus, having a way to sort and group those images would be helpful. At least, it is for me :-)
Added features:
- A new option in Tools/Sort for both Sort and Group By called Black-Pixels.
- Sort feature: calculates the percentage of black pixels in each images and sort them from 0 to 100%.
- Group By feature: Uses the Bins slider number selected and creates 100/Bins folders with the images percentage of black pixels. ie. a Bins of 5 will create 20 folders with the first one containing images with 0-5% of black pixels, the second one with 6-10%, etc. A static method as been added for the rounding error.
* PEP8 fixes
Co-authored-by: Olivier Gagnon <ogagnon@ludia.com>
- Add support for different mask centering
- Update legacy alignments to store mask centering
- Bugfix: lib.image ImageIO. Ensure unique queues are created (fixes mask tool when Face is input and an output folder is provided)
