Skip to content

Latest commit

 

History

History
12 lines (9 loc) · 792 Bytes

README.md

File metadata and controls

12 lines (9 loc) · 792 Bytes

Sequential_Yoga_Pose_Assessment_for_Self_Training

A deep learning based approach for sequential assessment of Surya Namaskar yoga poses.

Methods utilised for developing the GUI for self training:

  1. 2D pose estimation
  2. 3D pose estimation
  3. Pose classification using K-Nearest Neighbours
  4. Featuring a timer to keep track of pose duration and sequences
  5. Developig indicators to show wrong limb positions (Geometry-based pose analysis)

For a detailed overview of the methods, network architecure used, pose classifier model used, logic behind pose comparison and feedback comments kindly refer to this report.