kNOw SIGN Source Code Documentation

Introduction

kNOw SIGN is a sign language (ASL) recognition project that utilizes real-time and efficient programming to help mute and deaf by establishing two-way communication channel with normal people who have never studied sign language.

Source Code documentation is explained here, for more information about the project please visit kNOw SIGN

Project Structure

The project has 3 main parts to achieve the recognition process:

1. The hand detection part where all the frame pre-processing needed for recognition is done.
2. The recognition part which is responsible for using the deep learning model to inference the letters recognized by hand gestures.
3. The display section (Raspberry Pi version only) which handels the output of the recognition resuls to the OLED dispaly

Processing Stages

The project involves several stages to produce the recognition results and each stage is done but one or more functions of the program. These stages are done by:

1. FrameGen
2. handDetect
3. InterHand
4. My_ROI
5. LetterRecog
6. LedDisplay

The description of these stages is explained in detailes in the individual class documentation.

Resources

• Github
• kNOw SIGN Page