| The paper discussed the basic knowledge of image compression, the image coding theory, the wavelet transform used for embedded coding, the Karhunen-Loeve transform used for removing the relativity in an image and the basic knowledge of the embedded coding. We know that it is difficult to further decorrelate the three components of a big color image using direct KLT. To resolve the problem, in this paper, we group a big color image firstly, then Karhunen-Loeve transform is applied to the color components of each group. Based on the magnitude distribution law of DWT coefficients of color image data in Karhunen-Loeve space, SPIHT algorithm is applied to the three KLT components of each group. For each specific threshold, only the coefficients equal to or larger than the threshold will be encoded. This can save many bits and improve coding performance especially at large number of group. |
PDF Full Text Request