Algorithmic Based VLSI Architecture of Integrated Image Compression for CMOS Image Sensor

The challenging task of a CMOS image sensor is to have a well robust compression processor for multimedia and medical applications because it requires huge storage requirement of image data like small bowel images, retinal images and mammograms from different modalities such as ultra sonography, magnetic resonance imaging, computed tomography and media files. To overcome image degradation at high compression ratio (CR), we have proposed an integrated compression technique which combines quadrant tree decomposition (QTD) and lifting based discrete wavelet transform (DWT) to provide promising results with high compression at low bit rates without losing any information of the image and the advancement in technology such as backside illumination makes CMOS sensor to use in medical imaging application. The input image is undergone QTD coding followed by lifting based DWT. The quantised and encoded transform coefficients are then decoded using inverse lifting based DWT to obtain the reconstructed image. Simulation results of our proposed compression scheme gives 80 % pixel level memory reduction at a PSNR around 40 dB and less number of resources are utilized for different functions implemented in Xilinx ISE 9.1i varies from 0 to 53 % and it finds application in the field of e-health, tele-consultation, tele-radiology, tele-matics and telemedicine.