Rapid expression recognition and application method based on face key points and geometric deformation

Abstract

The invention discloses a rapid expression recognition method based on face key points and geometric deformation, and the method comprises the following steps: collecting an image, and carrying out the face detection of an original image; positioning 68 face key points of the target face; taking a connecting line of the center of an inner eye corner and a nose tip as a vertical direction of a human face coordinate space, and calculating a feature vector and a distance capable of representing a mouth deformation degree from the key points; performing statistics on mouth shape deformation parameters of conventional and extreme expressions, performing mouth shape state judgment according to the obtained vector and distance, and querying an expression definition table to realize expression recognition. The facial expression recognition method based on the geometric deformation degree of the key points of the human face is suitable for being prevented from being constrained by a limited facial expression data set, and is particularly suitable for use scenes with strict distinction between exaggeration degrees and facial expression variety richness of similar facial expressions.

Description

technical field [0001] The invention relates to the fields of artificial intelligence and graphics and imaging, in particular to a method for fast expression recognition and application based on key points and geometric deformation of human faces. Background technique [0002] Facial expression recognition technology is widely used in various fields of life, such as emotional monitoring of patients in medical treatment, expression matching of human faces and cartoon characters in entertainment games, etc. With the upsurge of deep learning in the field of computer vision, expression recognition has also become one of the hot challenges in the field of scientific research. [0003] Expression recognition usually consists of three stages: face detection, feature extraction, and expression classification. Among them, feature extraction is mainly divided into geometric features of face key points, and automatic dimensionality reduction features based on original high-dimensional...

AI Technical Summary

Problems solved by technology

[0005] To sum up, the current data-driven facial expression recognition technology has the following weaknesses: 1. The range of changes in expressions made by different people is different; 2. There is no data that contains more abundant expressions, resulting in the inability to identify those that are not in the database. 3. Once new expressions are added, the entire model must be retrained. This step is quite time-consuming, and as the types of expressions increase, the complexity of the machine learning model also needs to increase to avoid underfitting; 4 . Machine learning models with high complexity are often difficult to achieve better real-time performance; 5. Image-based end-to-end expression recognition is easily affected by the image acquisition environment such as lighting, etc.

The above situations limit the usage scenarios and accuracy of facial expression recognition

Images