AVWAC based Gradual transitions detection in presence of high Object-Camera Motions and Uneven Illuminations

Advanced Multimedia Technology had provided a new lease of life to current videos and made life amazing in the field of entertainment. On the other hand, special lightening effects, rapid object-camera motions and scene blending have posed significant challenges in content based video retrieval. These factors have made special effects in close consistency with the gradual transitions to fool even a robust system. We propose an efficient and accurate gradual detection system to detect fade out and fade in transitions using a simple approach. The average value of wavelet approximation component (AVWAC) at level 5 for all the scene frames is used to analyze the presence of fade out and fade in. The position of global maxima and global minima points using AVWAC in successive frames are used to analyze and find the origin and end frames of respective gradual transitions. A fade out is detected when a global maxima is followed by a global minima whereas fade in are result of occurrence of global minima followed by global maxima. A common threshold efficiently determines the occurance of gradual transition. When the minimum value of AVWAC exceeds the threshold the peakvalley algorithm finds the occurance of the gradual transition. The technique is computationally simple and faster since the existence of either of the gradual transitions is verified using an experimentally found threshold before the analysis. The method is effective in high illuminations, rapid object and camera motion videos and fast color changing scenes when tested with real videos.