Multiple-viewing-angle human face detecting method and device thereof under complex scene

Abstract

The invention relates to the field of video monitoring and discloses a multiple-viewing-angle human face detecting method and a device thereof under complex scene. In the invention, at least one cascaded branchless node classifier is arranged in a cascading vector tree structure detector for carrying out human face detection to a child window to be detected and before at least one branch node classifier. A cascading classifier is led in between branch node classifiers and is utilized to carry out effective step-by-step removal to non-face child windows on all branches, the complexity of the branch node classifiers is reduced, the quantity of the contained weak classifiers is reduced, and the detecting speed is improved while the detection rate is ensured.

Description

technical field [0001] The invention relates to the field of video monitoring, in particular to face detection technology in the field of video monitoring. Background technique [0002] In recent years, due to the needs of anti-terrorism, homeland security and social security, various countries in the world have increased investment in the field of security, and identification is a core issue of security. With the rapid development of computer vision and pattern recognition technology, the identification technology based on biometric identification has been developed rapidly and has been applied in many fields, showing its huge commercial value and social value more and more. Face recognition, as the most user-accepted, most natural and most intuitive visual biometric technology, its technical and application value is being highlighted. With the widespread use of face recognition in entry-exit government projects such as biometric electronic passports, face recognition tech...

AI Technical Summary

Problems solved by technology

[0010] 1. Since each node classifier in the vector tree is a strong classifier, the node classifier containing multiple child nodes is a branch node classifier, such as figure 1 The solid line circle in , because the branch node classifier is responsible for the purpose of removing non-face sub-windows and sub-window sub-windows of different views, so these branch node classifiers are often very complicated, including too many weak classifiers, thus reducing the detection speed

[0011] 2. There is no connection between each node classifier in the vector tree, and decisions are made independently, thus wasting the classification information of the tested sub-window on its parent node classifier

[0012] 3. The feature pool of Haar features and EOH features used in the training process is too large, often containing tens of thousands or even hundreds of thousands of features, and only thousands of strong classification ability are selected from such a huge feature pool. Features are a very time consuming job

Therefore, this method often takes several weeks or even longer to train a robust multi-view face detector.

Images