Face detection method under angle change based on regional progressive calibration network

Abstract

The invention discloses a face detection method under angle change based on a regional progressive calibration network. The method comprises the following steps that: (1) a feature extraction and fusion module extracts features for candidate region generation and face detection from a whole image, and fuses the output of an intermediate layer of each convolution block, so that a feature map contains richer geometrical information; (2) a region generation network module samples the original image, a neural network is utilized to give a confidence coefficient of whether each sampling region contains a human face and a rough region size regression value, and a high-quality human face candidate region primary is generated; and (3) the detection and region size regression is gradually executedon the human face candidate region by the progressive face detection module in combination with the feature and the human face candidate region, and the change range of the plane angle of the face isreduced at the same time, thereby realizing human face detection under any change of the plane angle. The method is used for solving the problem of face detection under plane angle changes, features can be better extracted when the network parameter quantity is low, and the method is more suitable for small-scale face detection.

Description

Technical field [0001] The invention belongs to the technical field of face detection, and specifically relates to a face detection method based on a regional progressive calibration network under angle changes. Background technique [0002] As a specific scene application of target detection, face detection is one of the most basic and challenging problems in computer vision. A high-performance face detector is a key guarantee for face-related research, such as: face alignment, face recognition, age recognition, expression recognition, and behavior detection. Benefit from deep learning on targets and face detectors Success, the face detector has matured after years of development, and its application scenarios have gradually become more complex. When the face detector is applied to face detection with an excessively large plane rotation angle, face detection due to huge changes in the face It is very challenging to achieve high performance for the detector. For face detection u...

AI Technical Summary

Problems solved by technology

[0006] In view of this, the present invention provides a face detection method based on a regional progressive calibration network under angle changes, which is used to solve the face detection problem under plane angle changes, and can better extract features when the network parameters are low. It is more suitable for small-scale face detection

Images