@Santosh: It depends on the type of the image. An image cannot be represented in a (lossy) compressed format using fourier or wavelet coefficients if we are to detect modifications that are at the pixel level.
For example: No compression technique will be able to compress an image with pixels sampled from a perfectly random source. Hence, in the worst case - we will have to store each and every pixel of the image. Improvements are only possible if we are given that such images are rare or that the data in the images is compressible or that we are allowed a certain degree of pixel error tolerance. -- DK http://twitter.com/divyekapoor http://www.divye.in -- You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To view this discussion on the web visit https://groups.google.com/d/msg/algogeeks/-/LeY_7ttPal0J. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/algogeeks?hl=en.
