Lossless CFA image compression algorithm for wireless capsule endoscopy

This paper presents hardware-oriented lossless color filter array (CFA) image compression algorithm. Presented algorithm and hardware implementation in Verilog Hardware Description Language are based on existing solutions with additional elements increasing the efficiency of the algorithms. In particular, the new method of parameter values calculation for the Golomb-Rice encoding and the zero run-length encoding has been added. HDL code was written without any primitives, so it can be synthesized independently for each platform. In exchange for a slight increase in computational complexity, it was possible to reduce required bit rate by an average of 7.2% when testing on a set containing endoscopy images. This paper presents elements of other work on which the algorithm is based and characterizes specific optimizations and their impact on the final algorithm result. Finally, these results will be compared with other available solutions.