Mouth shape image recognition system based on machine learning and method of identifying sound production through face line

Abstract

The invention discloses a mouth shape image recognition system based on machine learning. The mouth shape image recognition system based on machine learning comprises a mouth shape video image acquisition module, a video analysis module, a main control module, a mouth shape model matching module and a voice output module. The mouth shape video image acquisition module, the video analysis module, the main control module, the mouth shape model matching module and the voice output module are electrically connected in sequence. The system further comprises a mouth shape model library module, whichis electrically connected with the main control module. The mouth shape video image acquisition module acquires a lip video image of a target person through a front camera arranged in the image acquisition module. The invention also provides a method of identifying sound production through face lines. According to the invention, lip language is converted into image information through the camera,and information processing and conversion can be rapidly carried out. The system has a deep learning function, can be gradually adapted to expression habits of a user, and has certain real-time performance and practicability.

Description

technical field [0001] The invention relates to a system for recognizing the sound produced by lip language, especially suitable for deaf-mute people, belongs to the field of electronic technology, and specifically relates to a machine learning-based mouth shape image recognition system. The invention also provides a face recognition system. Pattern recognition method of vocalization. Background technique [0002] Most of the current deaf-mute interactive devices are realized through information collection and comparison of sign language images. mobile phone" patent. It can only be input through text, and cannot read the sound through lip language, which greatly limits the practicability. Poor real-time performance, not easy to use. Contents of the invention [0003] In order to solve the problems existing in the prior art, the present invention provides a mouth shape image recognition system that can directly make the sound that others want to make through the collecti...

AI Technical Summary

Problems solved by technology

It can only be input through text, and cannot read the sound through lip language, which greatly limits the practicality

Poor real-time performance, not convenient to use

Images