i'm doing my project in image denoising. i need to calculate inter and intra pixel dependency between subbands and caculate threshold with respect to it.
Image compression is a widely targeted investigated area. Many compression standards are in place. But still here there is room for high compression with the rebuilding of quality. The JPEG standard uses the Discrete Cosine Transformation (DCT) for compression. In this article, we present the new method of using DCT, which is image compression by pixel correlation and its comparison with existing algorithms. Reconstruction of the image without any loss with high compression ratio has to be defined in this context. This algorithm uses only JPEG color images. In this component R, G and B of the color image are converted to YCbCr before applying the DCT transform. Y is luminance component; Cb and Cr are components of chrominance of the image. Three different images are required in different sizes for comparison with Huffman coding and arithmetic coding. The results are analyzed on the basis of the compression ratio.
The demonstrative images have correlation between spatially periodic pixels, because the interpolation strategies at any location of logically equivalent spatial pixels are identical. Taking this statistical characteristic, much research on forensic imaging has been done recently. We proposed a generalized neural network framework to simulate stylized computational rules in demosaicking through bias and adjustment of weight value. As the experiments demonstrate, our framework is effective in recognizing demosaicking algorithms for raw CFA images, as well as digital photo authentication, as compared to state-of-the-art methods.