In this project various gestures and standard sign languages were recognized using OpenCV and Mediapipe libraries of Machine Learning recognition algorithm and they were implemented on the Raspberry Pi's Raspbian OS.
◾ The main focus of this project is to provide ease for the verbally mute and hearing-impared people to communicate fluently in their daily chores. This is the base system that can be further expanded to train the standard sign language using raspberry pi.
◾ This system trained a total of 20 standard gestures namely - numbers (from one to nine and zero), just a part, approval, three finger salute, you lost to me, direction, ok, like, play, no, good luck, and done. There are two options provided for user - 1 to train the data and 0 to recognize the data. We implemented this project on raspberry pi 3 by installing the mediapipe and opencv libraries.
◾ It can be implemented using either Thonny IDE or Jupyter Notebook (preferable) in Raspberry Pi.
This repository contains various sections -
- gesture_recog_windows-os.py - This file is the python source code for the gesture recognition system tested on Jupyter Notebook of Windows OS.
- gesture_recog_raspberrypi-os.py - This is the python source code for the gesture recognition system tested on Jupyter Notebook of Raspberry Pi 3 OS i.e. Raspbian OS.
- gesture_recognition.pkl - This file is the training data for the standard 20 gestures that we've trained during the implementation of our project.
- Gesture Recognition Signs - It is an issue providing the glimpses of the gestures that we've trained. All 20 gestures were recognized and captured by the system.
◻ Source - https://toptechboy.com/category/artificial-intelligence/