Real-Time Sign Language Recognition for enhanced communication

Real-Time Sign Language Recognition System

• Implemented SVM and MobileNetV1 on static image datasets for sign classification.

• Developed EfficientNet-B0 + LSTM architecture for sequence-based image inputs and LSTM on video

streams.

• Created a custom dataset of ARSL signs (webcam recordings + manual labeling) and integrated ASL data to

build recognition pipelines.

• Designed an enhanced two-way communication interface facilitating seamless interaction between deaf and

hearing users.