基于帧前景的特征距离判据的镜头边界检测
Shot Boundary Detection Based On Feature Distance Criterion of The Frame foreground Area.
提出了一种简单有效的基于HSV空间镜头边界检测方法，本文综合考虑了视频帧全图像像素点与局部颜色直方图特征。首先，通过对视频帧全图的像素点进行运算提取每一帧的有效前景运动区域，然后提取该区域的颜色直方图。利用滑动窗口计算当前帧的前后两组视频帧的颜色直方图类间与类内距离，构造有效颜色特征的距离判据进行镜头边界检测。实验结果表明对镜头切变与渐变有良好的检测能力。
Proposed a simple and effective shot boundary detection method based on HSV space, this paper comprehensive consideration of the whole image pixel video frame and the local color histogram feature. First, it operates the whole pixel map for exacting the effective foreground movement area, and then exacting the color histogram in the region, finally, the sliding window around the current frame is adopted to compute the between-class distance and within-class distance of the color histogram, constructing effective color features of the distance criterion for shot boundary detection. Experimental results show a better detectability of the cut transition and gradual transition.
段豪、杨倩、雷少帅、谢刚
自动化技术、自动化技术设备计算技术、计算机技术
镜头边界检测HSV空间帧前景运动区域距离可分性判据自适应阈值
Shot boundary detectionHSV spaceFrame foreground movement areaDistance separability criterionLocal adaptive threshold
段豪,杨倩,雷少帅,谢刚.基于帧前景的特征距离判据的镜头边界检测[EB/OL].(2011-09-15)[2025-08-16].http://www.paper.edu.cn/releasepaper/content/201109-217.点此复制
评论