METHOD FOR DETECTING CAPTION SYNTHETIC KEY FRAME IN VIDEO STREAM

摘要

PURPOSE: A method of extracting a video text composite key frame is provided to efficiently summarize video contents, search and filter a desired portion of the video contents using the summarized result in a multimedia summarizing and browsing system. CONSTITUTION: Significance weights with respect to detected text regions are allocated to text elements, respectively. Texts to be synthesized are selected from the video text based on the weights. The selected texts are synthesized into one image to set the image as a text composite key frame representing a specific video section. The text elements include the size of a text region, the size of the average text element in the text region, and a display duration of the text region.