Skip to content

Latest commit

 

History

History
20 lines (14 loc) · 623 Bytes

README.md

File metadata and controls

20 lines (14 loc) · 623 Bytes
Raw

Fingerspelling PoseNet: Enhancing Fingerspelling Translation with Pose-Based Transformer Models:

alt text

Paper Link: Arxiv

  1. Download the Chicago Fingerspelling dataset from Here

  2. Extract Pose using Media-Pipe 2.1. Install Media-Pipe Here

  3. Run Media-Pipe on input data using extract_pose_mediapipe.py

python extract_pose_mediapipe.py
  1. Run
python train_with_lm.py