Mostafa Atahar, Khan Tareq, Wahid Khan
Annu Int Conf IEEE Eng Med Biol Soc. 2014;2014:2452-5. doi: 10.1109/EMBC.2014.6944118.
Video capsule endoscopy is a non-invasive technique to receive images of intestine for medical diagnostics. The main design challenges of endoscopy capsule are accruing and transmitting acceptable quality images by utilizing as less hardware and battery power as possible. In order to save wireless transmission power and bandwidth, an efficient image compression algorithm needs to be implemented inside the endoscopy electronic capsule. In this paper, an integer discrete-cosine-transform (DCT) based algorithm is presented that works on a low-complexity color-space specially designed for wireless capsule endoscopy application. First of all, thousands of human endoscopic images and video frames have been analyzed to identify special intestinal features present in those frames. Then a color space, referred as YEF, is used. The YEF converter is lossless and takes only a few adders and shift operation to implement. A low-cost quantization scheme with variable chroma sub-sampling options is also implemented to achieve higher compression. Comparing with the existing works, the proposed transform coding based compressor performs strongly with an average compression ratio of 85% and a high image quality index, peak-signal-to-noise ratio (PSNR) of 52 dB.
视频胶囊内镜检查是一种用于医学诊断的获取肠道图像的非侵入性技术。内镜胶囊的主要设计挑战在于，要通过尽可能少地使用硬件和电池电量来采集和传输质量合格的图像。为了节省无线传输功率和带宽，需要在内窥镜电子胶囊内部实现一种高效的图像压缩算法。本文提出了一种基于整数离散余弦变换（DCT）的算法，该算法在专门为无线胶囊内镜应用设计的低复杂度颜色空间上运行。首先，分析了数千张人体内镜图像和视频帧，以识别这些帧中存在的特殊肠道特征。然后使用了一种称为YEF的颜色空间。YEF转换器是无损的，并且只需要几个加法器和移位操作即可实现。还实现了一种具有可变色度子采样选项的低成本量化方案，以实现更高的压缩率。与现有工作相比，所提出的基于变换编码的压缩器表现出色，平均压缩率为85%，图像质量指标峰值信噪比（PSNR）为52dB。
