Abstract: Method disclosed in the present invention is proposed to detect and eliminate flash scene in digital video. Particularly, averaging shot distribution of digital video is employed as knowledge to develop the algorithm of the present invention, and which is incorporated and used to identify three general types of shot distribution for flash scene event. In the present invention, the luminance difference between two consecutive frames is instead of actually analyzing the visual content so as to reduce the computational complexity. As a result, positions of flash frames can be exactly detected from the video signal for many applications. The method comprises the steps: frames are extracted from a video sequence inputted. A luminance difference by two adjoining frames is calculated. Then a histogram is made to record the differences, and a threshold is determined accordingly. Then the flash scene is detected and categorized into three types, and finally being eliminated.

Abstract: A sports video retrieval method involves segregation of the video into multiple segments of different indexing conditions based on audience cheers recorded; the exciting intensity of the cheers being scored as the audio level of the cheers being in proportion of the wonderful performance of the athletes; cheer segments being retrieved and given rank depending on the score to each segment to provide the sports video segmental indexing with simple audio even including contents of video features.

Abstract: A method to detect anchorperson segment in news reporting by using visual characteristics to provide the basis to divide news into various categories includes steps of providing news image for skin color detection on the image with color space; applying morphology depending on whether the object in the image subject to skin color detection is moving to eliminate noise surrounding the image of the face to solve the region of the face of the anchorperson; and performing anchorperson detection once again by detecting the probable anchorperson segment.

Abstract: A video segmentation method for segmenting video clips according to content of the video clips is disclosed. The method comprises scanning pixels of video frames with a first horizontal scan line to determine if colors of the pixels fall within a predetermined color range; creating a color map utilizing pixels located on the first horizontal scan line from a plurality of successive video frames; labeling the current video segment as a candidate video segment if the color map indicates the presence of a stable region of pixels falling within the predetermined color range for a predetermined number of successive video frames; and performing histogram color comparisons on the stable regions for detecting shot transitions. Audio signals of the video clips may also be analyzed to further verify the candidate video segments.