realtime-sign-language-recognition's Issues
Implement multithreading code to process I/O into separate threads to increase the program responsiveness
Detect if a hand is insight to remove the background and start the recognition
Fix the TensorFlow Lite build error with Raspberry Pi
Inference model need more training and bigger dataset to provide better recognition
Use callbacks and event-driven approach to achieve better performance
optimize the code for the final release
Create and run Github page for kNOw SIGN
integrate OpenCV code with Raspberry Pi camera
- Create content for weekly post
- Design weekly post graphics
Create unit tests for the different parts of the program and create executable for the tests
Repo Readme need to be updated with the current progress
- add openCV installation guide
- add hardware configuration
- add model training methodology
Background removal code in kNOw SIGN-pc branch is not working. Black screen appears after running and the detection is incorrect.
The background need to be removed from the camera live stream in ordered to improve the accuracy.
Create the display driver for RPi and implement the recognition output class