PRE-PROCESSING AND VECTOR QUANTIZATION BASED APPROACH FOR CFA DATA COMPRESSION IN WIRELESS ENDOSCOPY CAPSULE

In wireless endoscopy capsule, an efficient, low-complexity compression approach for Bayer CFA data is critical for the low power design of the entire system. In this paper, a compression approach based on pre-processing and vector quantization is proposed. The CFA raw data are first low pass filtered during pre-processing. Then, pairs of pixels are vector quantized into macros of 9 bits by applying block partition and code mapping in succession. After rearranging, these macros are entropy compressed by JPEG-LS. By control of the pre-processor, both near-lossless and lossy compression can be realized. The effectiveness of our block partition scheme has been demonstrated by statistical experiments. Simulation results show that the proposed approach has a good performance in compression rate as well as reconstructed quality