49 lines (32 loc) · 1.8 KB

Chengelog

Template and more info at https://keepachangelog.com/en/1.0.0/.

Types of changes:

Added for new features.
Changed for changes in existing functionality.
Deprecated for soon-to-be removed features.
Removed for now removed features.
Fixed for any bug fixes.
Security in case of vulnerabilities.

[Planned]

[Unreleased]

Changed

[0.4] - 2025-08-20

Switched from old mediapipe 0.8 FaceMesh solution to new 0.10 FaceLandmarker.
- Number of landmasks increased from 468 to 478!Thus, the output shape of the landmark infomation numpy file is changed to [N][478][3].
Removed cv2 dependency from face extraction.
- Now OpenCV is needed only for the computation of the motion energy.
Added code to extract blendshape data from faces. 52 blendshapes are optionally saved into a numyp array of shape [N][52]. See option --outblendshapes in extract_face_data module.

[0.3] - 2025-08-14

Switch to MediaPipe 0.10.x (from 0.8.9: API changes)
Update MTCNN to 1.0.0
Compatible with Python 3.11

[0.2] - 2025-08-12

Renamed package from dfki_sl_videotools to slvideotools
Added more methods for face detection: in script extract_face_bounds, parameter --method can be either mediapipe or mtcnn
Added a script to extract video motion energy using Optical Flow.
Removed option --outvideo from the face detection script. Bounds can be overlayed in dedicated script draw_bbox
Framework to "produce" and "consume" frames from either videos or directories of images

[0.1] - 2022-01-26

First release with command lines for face movement detection, crop, trim, facemesh extraction
Added this Changelog :-)
Face extraction tool can save save landmarks and includes a "normalization" option to transform the landmark to a camera-facing neutral position.