| Summary: | 碩士 === 國立中正大學 === 資訊工程研究所 === 88 === In this thesis, we propose an edge detection scheme for color images. Besides, there are also two applications of the edge detection method respectively to the fields image compression and copyright protection. As for the edge detection scheme of color images, principal component analysis (PCA) is adopted to reduce the color dimensions from three to one; also, the two-level quadtree segmentation is employed to save the computation time. As for image compression, on the other hand, the blocks of the image is separated into two categories at first. They are smooth blocks and edge blocks. Then two different methods are employed to compress the smooth blocks and edge blocks, respectively. As for copyright protection, the author uses the concept of moment-preserving to extract the characteristics of intensity of images. Then this extracted image feature and the owner’s watermark are employed to generate the secret key for copyright detection.
|
|---|
