9th International Conference on ICT for Sustainable Development

Authors - Gaurav Boob, Somya Maheshwari, Abhishek Jain, R Sreemathy
Abstract - Delving into the realm of automated lip-reading systems, this study conducts a comprehensive evaluation, centering its examination on pivotal components such as audio-visual datasets, feature extraction techniques, classification networks, and frameworks. Through this analysis, significant gaps and shortcomings in existing systems are uncovered. Initially, the study aims to assess the advantages of various architectural approaches, including Attention- Transformers, Temporal Convolutional Networks (TCNs), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU). Additionally, it reviews the most recent lip-reading devices available on the market as of early 2023. The goal of this survey is to inform and shape future research in automated lip-reading by shedding light on current approaches.