A Deep Learning-Based Lip Reading Module Applied for Forensic Analysis and Silent Speech Decoding

Machine learning and Deep learning algorithms in digital forensics play a significant role in crime detection aiding forensic analysis through identifying patterns, trends based on data. Audio and Video data serve as a great electronic evidence in digital forensics for investigation purpose. In forensic investigations, crucial video footage often lacks audio or suffers from poor sound quality, making it difficult for analysts to fully grasp the conversations that took place and traditional speech analysis methods depend on audio data, leaving a significant void when only silent footage is available. Several existing systems struggle with contextual understanding, that can lead to misinterpretations, and also raise privacy and ethical issues when applied in surveillance and forensic contexts. Thus chapter proposes to provide a solution employing advanced computer vision and deep learning methods called Lip reading module for forensic analysis and silent speech decoding that seeks to fill this void by offering a way to extract speech information from lip movements, potentially revealing insights that would otherwise remain hidden. The modular design of the project makes it a potential forensic analysis assistant tool for silent speech decoding.