A Partially Occluded Face Recognition Method Based on Data Augmentation

Abstract

The invention discloses a partially occluded face recognition method based on data enhancement, which uses artificial intelligence deep convolutional neural network to perform face alignment, facial features detection, face feature extraction, and face matching multiple times; through data enhancement Based on the principle of pre-judging the occlusion type of the standard face occluder, pre-judging the identity of the occluded face, and pre-judging the occlusion type of the standard face occluder, the two pre-judgment results are integrated for data enhancement. Based on the face, the occluded face database is reconstructed, and finally the original occluded face is matched with the reconstructed occluded face database. This technology solves the technical problems of the existing occluded face recognition method that the feature information is lost due to occlusion, the difficulty of face recognition is increased, and the recognition accuracy is reduced, and the technical effect of improving the accuracy and efficiency of occluded face recognition is achieved.

Description

technical field [0001] The present invention relates to the technical field of face recognition, in particular to a partially occluded face recognition method based on data enhancement. Background technique [0002] As an important machine vision technology, face recognition plays a major role in the field of artificial intelligence. In the actual process of face image processing, the occlusion of the face image, especially the occlusion of important features (such as eye occlusion caused by sunglasses, mouth occlusion caused by masks) occurs frequently. It is difficult to extract true and accurate feature descriptions from occluded face areas. The inaccuracy of feature descriptions greatly increases the difficulty of face recognition technology. If traditional machine vision recognition methods are used, the recognition accuracy will inevitably drop significantly. Face recognition technology needs to face technical difficulties such as occlusion, lighting, expression, post...

AI Technical Summary

Problems solved by technology

[0014] The present invention provides a partially occluded face recognition method based on data enhancement, which solves the problems in the existing occluded face recognition methods that the feature information is lost due to occlusion, the difficulty of face recognition increases, the recognition accuracy is low, and the recognition speed is slow technical problems, and achieved the technical effect of improving the accuracy and efficiency of occluded face recognition

Images