About

PSG-CAVSR is the Centre for Audio Visual Speech recognition developed under the Department of Science and Technology - Interdisciplinary Cyber Physical Systems (ICPS). The centre hosts an intelligent lecture capturing system with dedicated High Performance Computing GPU server for intensive research. It aims to develop a synergistic relationship with the leaders of industry, academia and government to provide opportunity for the students to work in deep learning. It is a centre for high performance computing, video processing, speech analytics and research using deep learning.

Objectives

• To develop an audio-video recognition system capable of efficiently capturing the lecture delivered by the teacher.
• To create online educational video library.
• To convert the lecture content into readable lecture notes in English.
• To convert the English lecture notes into native language.
• To create a repository of lecture content (in English and native language) for the student community.

Novelty in delivery of the innovations

• Using ASVR for learning by hearing imparied.
• Increasing speech to text accuracy using word prediction.
• Combining speech with figures/diagrams projected through Digital Lectern.
• E-accessibility providing lecture content as text accessible outside the classroom.
• Promoting barrie-free education.

Expected output

• A small class room setup to capture audio and video.
• Speech API to recognize speech from audio.
• Lip reading API to read lips.
• Word prediction API to predict next word.
• Subtitled video lecture.
• Textual lecture content in English and Tamil.
• Commercialization of the AVSR.
• Applying for parents on research outcomes.

Facilities Available in CAVSR

GPU Server

High Performance Workstation

Interactive Display

Camera

Body Pack Microphone

Audio Mixer

Video Recorder

Video Capture Card

Loud Speaker

Studio Monitor

Digital lectern

Teleprompter

Document Camera

×