Deep Learning Techniques for Enhancing Lip-Reading and Speech-Reading: A Survey

Deep Learning Techniques for Enhancing Lip-Reading and Speech-Reading: A Survey

Akshay N Patel                                                 Prof. S.K. Hadia

Student(GTU-GSET)                                         Associate Professor(GTU-GSET)

akpatel3290@gmail.com                                      asso_s­_k_hadia@gtu.edu.in

Abstract:

In our noisy world, understanding spoken words can be challenging. This survey explores how deep learning can unlock the language of lips, making it accessible to a broader audience. Aimed not only at technical enthusiasts but also at those wishing to decipher whispered conversations, it delves into the applications of deep learning for lip-reading and speech-reading. The focus is on practical uses, such as aiding individuals with hearing difficulties and improving communication in noisy settings like streets or classrooms. This survey provides a user-friendly journey through the advancements in this field, empowering researchers and developers to utilize deep learning. By making silent voices heard, it envisions a future where understanding conversations, even in the middle of noise, becomes a reality.

Keywords: Lip-reading, Hearing impairment, Silent communication, Feature extraction, Multimodal fusion, Deep learning