METHOD FOR ENHANCING BLIND DETECTION OF NON-SHIFT WATERMARK WITH FILTERING AND/OR DOWNSCALED DATA AS OBJECT

摘要

PROBLEM TO BE SOLVED: To provide a method for detecting watermark information from "downscaled" and filtered data. SOLUTION: Watermarking technology that efficiently resists shifting and cropping operations in real-time video exists. However it is not robust against downscaling and filtering. The existing watermark detection technique is improved so as to resist downscaling and filtering by using a "RAKE-SPOMF" algorithm to collect and re-combine weak watermark detection peaks dispersed by the downscaling and filtering. Guidelines are presented to illustrate how proper up-sampling of the downscaled video should be performed prior to a RAKE-SPOMF detection process. This method is effective in enhancing the robustness of blind watermark detection for watermarked video data that are subjected to downscaling, filtering, shifting, cropping and MPEG compression. Hence it is expected to be useful in real-time video watermarking applications such as copy protection for DVD and DVB. COPYRIGHT: (C)2004,JPO&NCIPI