A multiple transform based scheme for still color image compression

Still color images can be represented using different models, each of which has its complexity, and advantages. The commonly used models are the RGB and the YUV models. In the RGB model, the three primary colors are employed, while in the YUV and the HSI model, the luminance is separated from the chrominance. In this technical contribution, a novel technique is applied to represent and compress still color images. First, the image is represented using the YUV model. The 2-D Y matrix is represented employing the 2-D multi-transform technique with the mixed basis function set comprising of the DCT and Walsh transforms. The U and V 2-D matrices are first decimated. Next, the U and V subsampled matrices are represented using the 3-D multi-transform model employing the DCT and Walsh transforms, providing better compression than the commonly used DCT. This model provides higher compression due to the mixed transform technique as well as the 3-D model taking advantage of the U and V correlation.