Explore lip reading with AI as an assistive technology for the deaf and hard of hearing in this 40-minute conference talk from Strange Loop. Delve into the challenges faced by those with hearing impairments and the limitations of current accommodations. Discover how deep learning methods can revolutionize accessibility through lip reading technology. Learn about the training process, network architecture, and data preparation involved in developing this solution. Examine the benefits and challenges of speech recognition and lip reading, and understand the importance of context in improving accuracy. Watch a demonstration of the implementation and discuss its potential impact on making events and communication more inclusive for the hearing impaired.
Overview
Syllabus
Intro
Disability Accommodation Gaps
Invisible Disability
Misconceptions
Assistive Technology
Old School Solutions for the D&HOH
Assistive Augmentation Solutions
Ideal Characteristics
Training Process
Deep Neural Network
Speech recognition is not unisensory
Benefits of lip reading
Challenges of Speech Recognition
Overcoming Challenges
Long Short Term Memory
Network Architecture Chosen
Data Preparation
Our Training Results
Implementation Demo
Unknown words
Context is important
Taught by
Strange Loop Conference
