An approach for imprecise transmission of TIFF image files through congested real-time ATM networks

The paper presents an approach to pre-process and post-process TIFF image files in real-time applications for maintaining an acceptable image quality when these files are transmitted in a congested ATM network. Compared to existing techniques, the approach introduces a very low transmission and computational overhead, allows a higher cell-loss rate, and produces a less image degradation. In this approach, the preprocessor rearranges the content of image files, and then packs these fixed image files into ATM cells for transmission. The post-processor unpacks the cells, restores the original content orders, and if part of the data has been discarded, performs data recovery to reconstruct imprecise image files as close to their original versions as possible. By marking an appropriate number of cells, a user can specify an acceptable lower bound on the quality of the image that is delivered at the destination. This mates it possible to control and know which part of the image data can be discarded. The experimental results show that by using the proposed approach, the quality of the received/recovered images decreases more slowly than the proportional increase in network congestion