Facial Expression Recognition Method Based on Partially Occluded Images

Abstract

The invention discloses a facial expression recognition method based on partially occluded images, comprising the following steps: 1. Preprocessing human face images containing N types of expressions in a known type of human facial expression database; 2. Calculating test concentration The similarity between the sample to be tested and the training sample in the training set is obtained to obtain the same category as the sample to be tested and the nearest neighbor image; 3. Reconstruct the occluded part of the sample to be tested; 4. Extract the reconstructed sample to be tested and the PWLD features of the training samples in the training set; 5. Use the SVM classifier to classify and identify all the samples to be tested in the test set. The present invention adopts the method of image matching, which can effectively reconstruct the occluded part of the image, and avoids the problem of incomplete feature representation when only extracting the non-occluded part. In addition, the three-layer pyramid structure adopted by the present invention extracts the global and local features of the image. , which enhances the accuracy of feature representation.

Description

Technical field: [0001] The invention relates to image reconstruction and feature extraction, and belongs to the field of image processing, in particular to a facial expression recognition method based on partially occluded images. Background technique: [0002] Facial expression recognition has received extensive attention in human-computer interaction, intelligent information processing, etc., but most of the current research is carried out in a controlled environment, which is difficult to adapt to the complexity and variability of the external environment. Covered by glasses, scarves, masks and some random occluders, the recognition rate of facial expressions is greatly reduced. In recent years, research on facial expression recognition under occlusion has become an important research direction. Nowadays, some researchers dealing with facial expression recognition under occlusion try to reconstruct the texture and geometric features of the occlusion part, so as to elimi...

AI Technical Summary

Problems solved by technology

Although SpPCA overcomes the disadvantage that PCA does not distinguish the importance of different parts when expressing different expressions, for larger occlusions, the area of ​​the occlusion part may be divided into a smaller area separately, in this smaller area After calculating the eigenvalues ​​and eigenvectors, the reconstructed image still contains more occlusions, which will also cause the problem of low image recognition rate.

Images