A survey on Audio-Visual Speech Recognition System

Speech is the most effective and vital form of communication between people and their surroundings. It can recognize nondeterministic audio along with given preponderance among decisions with a near probability by using an approach that makes advantage of image processing skills for lip reading. The idea of continuing research into the creation of an automatic voice recognition system, which is a positive approach to humancomputer interaction, inspired certain researchers. However, in actual use, the existing framework’s effectiveness is adversely impacted by the existence of an audial noisy environment, which makes it difficult to recognize actual speech. Due to this, AudioVisual Speech Recognition (AVSR) was developed in order to overcome these difficulties by utilizing both audio and video feeds. This paper is being taken down in order to describe the fundamental design of the different AVSR as well as how it has changed over time in various research projects. It also further distorts our impression of future audio visual speech recognition architectural designs.