Storyboard construction using segmentation of MPEG encoded news video

Proposes an algorithm to construct the storyboard using segmentation of MPEG encoded Korean news video. To do this, we use the MPEG encoded and decoded data. Different from general video data, news video data have fixed flow. We use the specific method to segment the video data using the flow. We don´t parse the video data in the whole region but in each region (anchor person shots and five news shots), and also we use the encoded data to improve the processing time. Firstly, we detect anchor person shot using the luminance DC mean value of all I-frames. This provides us with the first and end frame of anchor person frames. As a result, anchor person shots and, live news shots parse in different methods, and the accuracy and processing time of parsing the data are improved. We can obtain about 90% accuracy in spite of reduction in the processing time