Image Coding by Multi-Step, Adaptive Flux Interpolation

This paper describes and discusses a new technique, the multi-step adaptive flux interpolation (MAFI) and its application to image data for coding. The output of MAFI, when applied to an image, is still in an image form but has a more uniform feature density. This is because the original image has been warped by removing those rows and columns which contain mostly redundant pixels. It is also greatly reduced in size and the side information is minimal. The MAFI output can be further compressed using conventional coders, making its compression ratio even higher. Because of its warped nature, the MAFI output's statistics are also more consistent with the properties assumed by block-based discrete cosine transform (DCT) models.