Optimizing Hand Region Detection in MediaPipe Holistic Full-Body Pose Estimation to Improve Accuracy and Avoid Downstream Errors
This work addresses a specific technical bottleneck for sign language processing applications, representing an incremental improvement.
The paper tackles a flaw in MediaPipe Holistic's hand region detection that reduces sign language recognition accuracy, proposing a data-driven enhancement with additional features that yields higher Intersection-over-Union scores.
This paper addresses a critical flaw in MediaPipe Holistic's hand Region of Interest (ROI) prediction, which struggles with non-ideal hand orientations, affecting sign language recognition accuracy. We propose a data-driven approach to enhance ROI estimation, leveraging an enriched feature set including additional hand keypoints and the z-dimension. Our results demonstrate better estimates, with higher Intersection-over-Union compared to the current method. Our code and optimizations are available at https://github.com/sign-language-processing/mediapipe-hand-crop-fix.