Multi-size fusion face detection method, device and storage medium

Abstract

The embodiment of the present invention provides a multi-size fusion face detection method, comprising: inputting the initial face picture into the backbone network of VGG16, obtaining the regular face feature atlas and the face feature atlas to be normalized; Each feature map in the described human face feature atlas to be normalized obtains a normalized face feature atlas; the conventional face feature atlas and the normalized face feature map set are merged to obtain the final human face feature map set. Face feature atlas; In the final face feature atlas, any adjacent two face feature maps are subjected to feature fusion to obtain a fused multi-size face feature map; the multi-size face feature map is carried out Classify and score each classification result, and the classification result with the highest score is used as the face detection result. The embodiment of the present invention also provides an active interaction device and a non-transitory readable storage medium, which are used to implement the method. The invention can reduce the number of zooming times of pictures, and realize a higher detection rate of human face pictures.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of pattern recognition, and in particular to a method, device and storage medium for multi-size fusion face detection. Background technique [0002] Most of today's face detection technologies use image pyramid technology to continuously scale the same image to be detected to deal with the multi-scale detection problems of faces, which will waste a lot of time on unnecessary calculations , resulting in a decrease in detection efficiency. Face detection with multi-size feature fusion detects faces on different size features. This reduces the number of times the image needs to be scaled and passed through the face detector. A higher detection rate can be achieved, and the detection speed is faster at the same time, which provides an effective inspiration for solving the problem of low detection efficiency. Therefore, how to develop a face detection method with multi-size features has b...

AI Technical Summary

Problems solved by technology

[0002] Most of today's face detection technologies use image pyramid technology to continuously scale the same image to be detected to deal with the multi-scale detection problems of faces, which will waste a lot of time on unnecessary calculations , resulting in a decrease in detection efficiency

Images