Real-Time Sign Language using Tensor Flow Ob...

Real-Time Sign Language using Tensor Flow Object Detection

Author Name : Karthikeyan Shanmugam

ABSTRACT The project entitled "Real-Time Sign Language Detection Using TensorFlow Object Detection" was developed to facilitate barrier-free communication for the hearing-impaired and speech-impaired population. Real-time capture of hand gesture using MediaPipe for hand pose estimation, and objects caught up were used for representing hand in visual snapshots which Random Forest classifier uses to process video input and classify hand gestures into the corresponding sign with the display of on-screen feedback. The main challenge the project would pose is to address accuracy, as well as complex backgrounds with robust preprocessing and efficient feature-based classification, which will further lead it to a promising future development like multi-sign recognition or personalized models use for improved inclusivity.