crosstyan
4a5cfde245
- Introduced the `OneEuroFilter` class to implement an adaptive low-pass filter for smoothing keypoint data, enhancing tracking stability during varying movement speeds. - Implemented methods for initialization, prediction, and updating of keypoints, allowing for dynamic adjustment of smoothing based on movement. - Added detailed documentation and type hints to clarify the filter's functionality and parameters. - Improved the handling of timestamps and filtering logic to ensure accurate predictions and updates.
jupytext --to py:percent <script>.ipynb