Face feature weight mapping method, face recognition method and device

Abstract

The invention discloses a face feature weight mapping method, a face recognition method and a face recognition device. The face feature weight mapping method comprises the following steps: acquiring a first feature vector of a normal face image and a second feature vector corresponding to a specific mask; calculating an offset vector according to the first feature vector and the second feature vector; performing iterative training according to the plurality of normal face images to obtain a plurality of offset vectors; and according to the plurality of offset vectors, calculating a feature weight vector of the specific mask, and generating a mapping relationship between the feature weight vector and the specific mask. According to the face feature weight mapping method disclosed by the invention, the difference of the covered part is converted into the offset information by calculating the offset vector, so that machine identification is facilitated. Through iterative training, errors can be reduced, and under-fitting can be prevented. Through the mapping relation, the feature weight of the covered part can be independently reduced when the image is covered, so that the condition that any position in the image is covered can be detected through a single database.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence image recognition, in particular to a face feature weight mapping method, a face recognition method and a device thereof. Background technique [0002] At present, there are generally two face recognition methods for wearing masks in the industry: [0003] Follow the normal face recognition steps, such as figure 1 As shown in Fig. 1, use the face recognition model to extract face features, and then compare the features with the face feature database to find out the face identification number with the largest similarity and greater than the set threshold. This method needs to add face data wearing masks in the offline learning phase of face recognition, which is compatible with the changes in facial features caused by wearing masks, but adding face data wearing masks to train the face recognition model weakens the The characteristics of the occluded part also indirectly weaken the...

AI Technical Summary

Problems solved by technology

[0005] Therefore, the technical problem to be solved by the present invention is to overcome the defects in the prior art that cannot rely on a single database to accurately identify faces wearing or not wearing a mask, thereby providing a face feature weight mapping method, a face recognition method and its device

Images