Face comparison preprocessing method

Abstract

The invention discloses a face comparison preprocessing method. Global 68 face key points are obtained through a neural network algorithm; then, according to the estimated scale and angle, geometric correction is carried out on the face image, the shielding conditions of several key local features on the human face are detected at the same time, if shielding is found, the affine transformation function is solved according to the local neighborhood points and the symmetric points, and radial frame replacement of the symmetric unshielded positions is carried out, so that the shielded problem ofthe local features of the human face can be optimized, the human face recognition is preprocessed, and the recognition precision is improved. Through face comparison preprocessing of face local feature detection and frame replacement, the defect that face information is interfered by factors such as illumination and shielding can be overcome, the face features are further perfected, and the face recognition precision is improved. The face comparison preprocessing method has the advantages of being capable of automatically perfecting face information and high in face recognition precision.

Description

technical field [0001] The invention relates to the technical field of face recognition, in particular to a face comparison preprocessing method with high recognition accuracy. Background technique [0002] With the development of society and the advancement of science and technology, more and more artificial intelligence technologies and products are applied to our daily life. Especially in the process of rapid urbanization today, urbanization presents a dense population, complex levels of personnel, and a large population, which brings many new social security management issues. Using intelligent technical means to effectively digitize and informatize urban management, rationally utilize social security management resources, and improve social security and people's quality of life have become one of the urgent needs of the current society. With the breakthrough and gradual landing of artificial intelligence technology, the application in the field of smart security has al...

AI Technical Summary

Problems solved by technology

Among them, the application of face recognition function is one of the biggest application requirements in the field of smart security, and its effect has been limited by the accuracy of face recognition. Due to factors such as illumination and occlusion of the face, the face information is lost and other problems. appear, which has a great impact on the accuracy of face recognition

Images