Sign Language Recognition and Conversion to Audio using Image Processing and Machine Learning

Sign Language Recognition and Conversion to Audio using Image Processing and Machine Learning

Pragathi H1, Roopa M S 2
1 (PG Scholar, Computer Science and Engineering Department, Dayananda Sagar College of Engineering, Bengaluru)

2(Associate Professor ,Computer Science and Engineering Department, Dayananda Sagar College of Engineering, Bengaluru)

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - Sign language is one of the oldest and most natural forms of language for communication. Sign language is used by speech-impaired persons for communication. For a regular individual, sign language interpretation is exceedingly challenging. This project effort proposes a real-time method for finger spelling-based American sign language using neural networks. The hand is first put through a Gaussian filter in this manner to get rid of the noise. A classifier is used to predict the class of the hand gestures from the filter’s output. It displays the textformatted output for a specific sign from that class. Python’s pyttsx3 module is used to convert text to speech.

Key Words: American sign language, neural networks,Gaussian filter,pyttsx3