Abstract: Lip reading, the capacity to understand spoken language by visually examining the motions of a speaker's lips, offers enormous potential to improve human-computer interaction and close communication gaps for the hearing-impaired. This project introduces "LipNet," a cutting-edge web application that leverages deep learning technology to enable real-time lip reading and automatic speech recognition. The application's core functionality is built upon a state-of-the-art deep neural network architecture, tailored specifically for lip reading tasks. The network is trained on extensive datasets of labelled video sequences, to ensure robustness and adaptability in diverse scenarios. LipNet offers a user-friendly web interface, allowing users to upload the video. The system rapidly processes the visual data, extracting facial landmarks and lip features with exceptional precision. Through a combination of convolutional and recurrent layers, the deep learning model transforms these visual cues into text representations of the spoken content. LipNet's high-performance architecture ensures reduced latency, making it suitable for real-time lip reading applications, facilitating instantaneous communication for the hearing-impaired. This web application serves as a stepping stone towards a more inclusive and accessible future, where technology fosters seamless understanding and connectivity between individuals, regardless of their auditory abilities.
Keywords: LipNet, Deep Learning, visual data.
Works Cited:
Ramya H, Sundararajan G, Kumaran M " LipNet: Bridging Communication Gaps through Real-time Lip Reading and Speech Recognition ", IJARCCE International Journal of Advanced Research in Computer and Communication Engineering, vol. 12, no. 9, pp. 100-104, 2023. Crossref https://doi.org/10.17148/IJARCCE.2023.12917
|
DOI:
10.17148/IJARCCE.2023.12917
