METHOD FOR DETECTING A LAYOUT REGION WITHIN A VIDEO IMAGE SO THAT PRIOR KNOWLEDGE ABOUT A LAYOUT TO BE DETECTED CAN NOT BE DEMANDED, AND A METHOD FOR GENERATING AN IMAGE OF A REDUCED SIZE

摘要

PURPOSE: A method for detecting a layout region within a video image, and a method for generating an image of a reduced size are provided to detect the layout region such as a logo and/or a score within the video image and generate the image of the reduced size from a source image.;CONSTITUTION: By using a visual attention model, a saliency map of a source image is generated(S1). One or more layout region within the source image is determined from the saliency map. Two points are searched within at least one predetermined saliency map region(S2). The two points are consistent with a maximum saliency value of the saliency map. The two points are spaced as much as at least p points. A window about a search region including two maximum saliency regions is regulated(S3).;COPYRIGHT KIPO 2010