v0.0.3
tinyhumansai/openhumanv0.0.3Mar 7, 2025by ShenhanQian
AI Summary
This release enhances camera calibration support and streamlines the rendering pipeline. It introduces functionality for cameras with off-center principal points and improves the FlameViewer by enabling texture map loading.
Key Highlights
- Support for cameras with principal points not located at the image center
- Simplified rendering functions and FlameViewer texture map loading
- Added temporal smoothness term for expressions
New Features
- Support for cameras with non-center principal points
- FlameViewer texture map loading capability
- Temporal smoothness term for expressions
Full Release Notes
**Major changes:** - Support cameras whose principal points are not at the image center. (https://github.com/ShenhanQian/VHAP/commit/34afb35bfe71a8b31d4a95fa41ef9b625bc1fc24) - Simplify rendering functions and allow FlameViewer to load a texture map. (https://github.com/ShenhanQian/VHAP/commit/d8ce4e97a38c85bb22cfee5dacced09d51723b36) - Add temporal smoothness term for the expression. Refract and simplify loss names. (https://github.com/ShenhanQian/VHAP/commit/1603303eb8c60bab3a8d8f47c03386f8dfffa04b)