Rapid face positioning method and system applied to driver fatigue detection

Abstract

The invention discloses a rapid face positioning method and system applied to driver fatigue detection. After face detection is performed on a driver image and a face rectangular frame is obtained, face tracking is performed on the driver image, and the timeliness of face positioning is improved with the cooperation of face detection and face tracking; in a face tracking process, a result of face tracking is monitored, if the currently tracked frame of the driver image is a front face, face tracking is continued, and if the currently tracked frame of the driver image is a non-front face, face re-positioning is performed so that the drift problem during face tracking in the non-front surface condition is solved; and during face re-positioning, whether the face rectangular frame including the face exists is detected in a preset range regarding a first central coordinate as the center in the driver image. According to the method and system, the detection range is greatly reduced compared with the overall region of the driver image, the detection range is more accurate, detection is performed at an interval of a preset frame number, the calculation amount of an algorithm is reduced, and the timeliness and the accuracy of face positioning are improved.

Description

technical field [0001] The invention relates to the technical field of computer vision processing, in particular to a fast face positioning method and system applied to driver fatigue detection. Background technique [0002] With the progress of society and the improvement of people's living standards, more and more families own private cars, and the traffic safety problems that follow are worrying. Among them, the problem of "fatigue driving" is particularly serious, which leads to a large number of vehicle collisions every year. and heavy casualties. Usually, the key to driver fatigue detection is to judge whether the driver is tired by detecting some features of the face. Therefore, it is of great significance to study the fast face localization method applied to driver fatigue detection. [0003] At present, in the main methods of face positioning, the method based on geometric features is only suitable for the detection of a single frontal face in a simple background, ...

AI Technical Summary

Problems solved by technology

[0003] At present, in the main methods of face positioning, the method based on geometric features is only suitable for the detection of a single frontal face in a simple background, and the detection effect is not ideal in complex lighting or background environments or in the case of occlusion. The method based on the skin color model Simple and fast, suitable for monotonous backgrounds, the detection effect is not ideal in complex lighting or complex backgrounds, the method based on statistical theory uses classifiers for face detection, this method has high robustness and accuracy, so it is more Reliable, it is the research hotspot and main method of face positioning in recent years. Among them, the Viola face detection algorithm has a high accuracy rate and can handle complex backgrounds, so it is the most widely used, but the calculation complexity is high. Only Viola face detection algorithm is used. The face detection algorithm will have missed detection and false detection, and its real-time performance cannot be guaranteed

Images