关键词: 图像处理, 视频监控, 前景提取, 高斯变差, 遗传算法

Abstract: Aiming at the difficult problem of unsupervised or automatic foreground extraction, an automatic foreground extraction method based on genetic and difference of Gaussian (GFO) is proposed. Firstly, Gaussian variation is used to extract the relative important regions in the image, which is defined as candidate seed foreground. Secondly, using the edge information of the original image and candidate seeds foreground, the contour of foreground object contour is generated according to connectivity and convex sphere principle, called star convex contour. Thirdly, the adaptive function is constructed, the seed foreground is selected, and the genetic mechanism of selection, crossover and mutation is used to obtain the accurate and valid final foreground. The experimental results of the Achanta database and multiple videos show that the performance of the GFO method is superior to the existing Automatic foreground extraction based on difference of Gaussian (FMDOG) method, and the recognition precision, recall rate and F-Beta index are better the extraction effect.

Key words: Keywords: image processing, video surveillance, foreground extraction, difference of Gaussian, genetic algorithm