Eurosensors edition:XXIV location:Linz date:5-8 September 2010
A dedicated image compression algorithm was developed for wireless capsule endoscopy (WCE). The proposed approach reduces the amount image data without compromising the image quality. This allows an increase in image throughput to improve the real-time vision capabilities of a WCE system. The algorithm is based on a normalized Haar-wavelet transformation, aiming at a hardware-optimized implementation in a low-power FPGA or telemetry ASIC for wireless capsule endoscopes. Taking advantage of the limited color content of endoscopic images, this solution uses the YCbCr color space combined with efficient recoding of the wavelet transformation data. In this way adequate compression ratios at decent image qualities can be achieved in a hardware efficient thus low power implementation, outperforming existing hardware implemented algorithms.