An Illuminated Face Recognition Method Based on Illumination Estimation

Abstract

The invention discloses an illumination face recognition method based on illumination estimation. The method is as follows: firstly, the initial illumination is estimated from the original face image according to the image imaging model, and then the light occlusion correction illumination and illumination fusion factor is calculated according to the initial illumination , build a lighting fusion model to realize light occlusion correction, then use the full variational model to smooth and filter the light after light occlusion correction to obtain the final illumination of the original face image, and finally use the illumination invariant extraction model to obtain the illumination invariant of the original image . The present invention can better deal with the shadow edge problem, extract the illumination invariant that contains rich facial detail features and is closer to the original human face intrinsic features, not only can satisfy the imaging principle of the image, the slowly changing characteristics of the illumination, but also adapt to complex lighting environment.

Description

technical field [0001] The invention relates to a face recognition method, in particular to an illumination face recognition method based on illumination estimation. Background technique [0002] The complex lighting problem has always been a research hotspot in face recognition technology. In order to solve this problem, scholars at home and abroad have proposed many methods. Such methods mainly include methods based on illumination estimation and methods based on high-frequency feature extraction. These methods essentially separate the illumination and illumination invariants of the face image on the assumption that the illumination changes slowly and the illumination invariant changes rapidly, and directly or indirectly extract the illumination invariant information from the face image. [0003] Although these methods have made significant progress in complex illumination face recognition, they still have certain limitations. On the one hand, in a complex environment, t...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the defects of the prior art, the object of the present invention is to provide an effective face image illumination estimation and illumination invariant extraction algorithm, that is, a complex illumination face recognition method based on effective light estimation, combined with image fusion and smoothing filter application The illumination invariant extraction model obtains the illumination invariant of the original image, which can better deal with the shadow edge problem, and extract the illumination invariant that contains rich facial details and is closer to the original face intrinsic

Images