Multi-face detecting and tracking method

Abstract

The invention provides a multi-face detecting and tracking method, belonging to the technical field of artificial intelligence. Firstly, according to multiple face characteristics in a monitoring system video, the method with the combination of Haar characteristic and an Adaboost classifier is used to detect faces, the concrete procedure comprises the following steps: (1) for a video continuous sequence, a Gaussian mixture model is used to extract moving foreground as a first class interest region, a region from the previous frame detection out of a motion region to a face as a center is taken as a second class interest region, and the face detection of the two interest regions is carried out, (2) a mean shift method is used to achieve multi-target tracking, and the adaptive updating and tracking of multiple faces of the same video are satisfied at the same time, (3) the multi-target tracking algorithm in the step (2) and the detection algorithm in the step (1) are combined, and a mixed multi-target tracking face detection algorithm is developed. According to the multi-face detecting and tracking method, the missed detection caused by face direction change or facial expression change is solved, the detection rate is raised, and the requirement of real-time performance by a monitoring system is satisfied.

Description

technical field [0001] The invention relates to the fields of computer vision and artificial intelligence, in particular to a multi-face detection and tracking method. Background technique [0002] In recent years, face detection algorithms have made great progress in the academic field, but there are still many problems to be overcome when applied to intelligent video surveillance systems. The detection rate and the detection time are often negatively correlated. Usually, the face detection algorithm that meets the detection rate requirements cannot meet the real-time requirements due to the high complexity of the algorithm, while some algorithms that meet the real-time requirements cannot meet the detection rate requirements. Require. At present, many studies are focusing on improving the detection rate and false detection rate of face detection algorithms. Algorithms are usually more and more complex and time-consuming. In order to apply face detection algorithms to inte...

AI Technical Summary

Problems solved by technology

The detection rate and the detection time are often negatively correlated. Usually, the face detection algorithm that meets the detection rate requirements cannot meet the real-time requirements due to the high complexity of the algorithm, while some algorithms that meet the real-time requirements cannot meet the detection rate requirements. Require

At present, many studies are focusing on improving the detection rate and false detection rate of face detection algorithms. Algorithms are usually more and more complex and time-consuming. In order to apply face detection algorithms to intelligent video surveillance systems, it is necessary to Select and improve the existing face detection algorithm to develop an algorithm that can meet the real-time requirements of the monitoring system while ensuring the detection rate requirements

[0003] Now there are face detection methods based on inherent feature models and face detection methods based on statistical classification ideas. model, and then extract the features of the corresponding area of ​​the detected image to compare with the reference model to detect the face. This method has a small amount of calculation and a fast detection speed. The disadvantage is that it cannot comprehensively and systematically describe the features of the face.

In comparison, the face detection method based on the idea of ​​statistical classification replaces manual selection with machine learning algorithms, imitates the human brain to train a classifier, and then uses the classifier to judge whether there is a face in the detected area. The face detection method of statistical classification can reflect the face features very comprehensively and significantly improve the detection rate. Its disadvantage is that the amount of calculation is large, and it takes a lot of time in both the training and learning process and the detection process. In practical applications Difficult to meet real-time requirements

However, to achieve the goal of extracting diverse faces from environments of different complexity, the difficulties that need to be faced are as follows: (1) For the same person, the face is a non-rigid object, and the same face has different expressions and turns. 1. The characteristics displayed under different lighting conditions change; (2) For different people, the skin color is different, and the glasses, beards and other appendages on the face will also cause visual differences; (3) Complex background and Noise interference, some complex background areas with features close to the face will cause false detection; and face detection in complex scenarios with multiple targets is more difficult

Images