Face detection method and system

Abstract

The invention relates to a face detection method and system. The method comprises the following steps: acquiring a to-be-detected original image; carrying out skin color detection and foreground detection on an original image to extract a suspected face area in the original image, and drawing a target box for the suspected face area; defining the size of a target frame, and screening the target frame by adopting a Sobel edge detection algorithm to filter out non-target frames in the target frame; sequentially performing R-Net operation and O-Net operation are on the filtered target frame to obtain a target frame including a real face region and five feature points in each of the target frames for indicating a face contour.According to the technical scheme, a traditional MTCNN network is improved, a skin color detection method and a foreground detection method are combined, a suspected face area is extracted, the operand is greatly reduced under the condition that the frame scale of anoriginal image is not reduced, the detection precision and the operation time are balanced, the cost of face detection equipment is reduced, and higher universality is provided.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a face detection method and system. Background technique [0002] The purpose of face detection is to find all faces in a picture. In the prior art, there are several commonly used methods for face detection: [0003] 1. The early face detection method is to use manual extraction features, train classifiers, and perform face detection. For example, the face detector that comes with the opencv source code is a face detection using haar features. The disadvantage of this type of method is that the detection effect is not ideal when the environment changes strongly, such as low light conditions and incomplete faces. [0004] 2. The face detection algorithm inherited from the general target detection algorithm. For example, use faster-RCNN to detect faces. The effect is good, and it can adapt to environmental changes and face incompleteness, but the time consumption is ...

AI Technical Summary

Problems solved by technology

[0007] 1. Recall rate: Difficulties in face classification caused by changes in the scale of faces in complex backgrounds and external factors such as illumination (in the CNN method, solving such problems is generally done by adding such samples for training)

[0008] 2. Time-consuming: time consumption caused by multi-scale position positioning of faces

[0022] The disadvantage of the existing face detection method is that it is impossible to balance the accuracy requirements of detection and the requirements of low calculation amount. The detection scheme with relatively high accuracy has a large amount of calculation; the detection scheme with small calculation amount is not accurate enough.

At present, the face detection based on MTCNN is the most lightweight and effective face detection algorithm, but even so, it can only achieve 1-2 frames per second on the ARM3519, which is difficult to achieve real-time, and the current improvement method is generally It is carried out on the basis of zooming pictures and detecting every other frame, so that the detection distance is reduced and the target is seriously lost.

Images