Implantable narrow band image compressor for capsule endoscopy

A low-complexity narrow band image compressor for wireless capsule endoscopy application is presented in this paper. The compression algorithm is based on a simplified RGB-YUV color space converter, differential pulse coded modulation (DPCM) followed by optimized Golomb-Rice coding. Based on the nature of the narrow band endoscopic images, several sub-sampling schemes on the chrominance components are applied. The compressor does not need any buffer memory and can be interfaced with image sensors which send pixels in raster-scan fashion. The proposed algorithm has compression ratio of 81% and high reconstruction peak-signal-to-noise-ratio , over 40 dB. The proposed compressor is implemented in 0.18μm CMOS technology and consumes 2K standard gates, 0.67 mm × 0.71 mm silicon area, and 42 μW of power when working at 2 frames-per-second.