Detection of Dissolve transitions in High Quality Videos Using Histogram based Normalized Correlation Coefficients and Adaptive Threshold

Abstract

Rapid enhancements in Multimedia tools and features day per day have made entertainment amazing and the quality visual effects have attracted each and every individual to watch these days’ videos. The fast-changing scenes, light effects and undistinguishable blending of diverse frames have created challenges to researchers in detecting gradual transitions. The state of art editing mechanism has made difficult to perceptually distinguish and application software’s to highlight transitions between various scenes in videos due to marvellous mixing techniques available today. The proposed work is aimed to detect gradual transitions in videos using correlation coefficients obtained through color histograms and adaptive thresholding mechanism. Other gradual transitions including fade out, fade in, cuts are eliminated successfully and dissolves are found from the acquired video frames. The characteristics of normalized correlation coefficient are studied carefully and dissolve are extracted out in a simple manner with low computational and time complexity. The experimental results obtained over 14 videos with 25 dissolves proved remarkable in detecting 22 dissolve transitions successfully under lightning effects and object motions. The detection accuracy is affected when the start and end frame of the transition have deceptive similarity,similar of poor contrast background and large object motion over the original frame. The performance is measured in terms of precision, recall and F-measure.