- Developed a Starting app that can detect only 3 action.
- The App is based on Pose Landmarks Extraction.
- Created a Dataset of 3 action i.e. “Hello”, “Thanks”, “I Love you”.
- Each action contains the 30 video stored as NumPy arrays
- The array contain the Coordinates of the all the Landmarks, especially Hands Landmarks
- I trained a Sequential LSTM model
Hello.mp4
Thank.you.mp4
I.Love.you.mp4
Action.Recognition.app.mp4
Action : Hello | Action : Thanks | Action : I love You |
---|---|---|