這星期突然對好久前做的研究好有興趣因手機越來越強，想把它在手機運行。特別找回一篇當年(1999)我在 San Jose, CA present 的文獻
Selective block assignment approach for robust digital image watermarking
Kam-Shing Ng, Lee Ming Cheng, City Univ. of Hong Kong (Hong Kong)
Abstract
Digital watermark is used to protect digital image against any illegal reproduction and tampering. In the selective block assignment process, the image is divided into N X N pixel blocks and each block is Discrete Cosine Transformed (DCT). The set of blocks will be then selectively chosen to encode the copyright message. Each selective block will be incremented by a value, in order to maintain the invisibility of the watermarking image, the incremented value should be within a range. The selection of blocks is based on measurement of the content. Depends on the amount of messages stored and the signal to noise ratio (SNR) of the resultant image required, a threshold is decided. In practice, the threshold will be set such that the duplicated message or an error correction mechanism can also be included in order to increase its robustness. The decoding process should be carried out by using the threshold values to get back the locations that have watermark information. Then the watermarked image is subtracted from the original image to obtain the secret data. Simulation results show that the watermarked image looks visually identical to the original and with an SNR of 44.7 dB for Lenna and with SNR 43 dB for airplane with size 256 X 256 pixels.好覺得這個領域就快出現 普遍性的用途。我亦會先在我的小學網站試用，相信能成一時佳話。
