Reversible compression techniques for high-resolution digital teleradiology
G. Kuduvalli, Rangaraj Mandayam Rangayyan · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 1990
The sheer amount of data involved in the fields of Picture Archival and Communication systems (PACS) and Teleradiology has created a growing interest in efficient compression of medical images. Diagnostic quality digitization of typical medical images such as chest x-rays and mammograms requires about 4Kx4K pixels digitized to 10 or 12 bits. In most medical applications any loss of quality in compression and decompression of images cannot be tolerated. We are investigating coding techniques for reversible compression of medical images in the above context. In this paper we present the results of application of adaptive linear predictive coding techniques to a few typical medical images digitized to about 4Kx4Kxl2-bits using an Eikonix 1412 camera. It was seen that about 75 reduction in transmission times is possible without any loss of image quality for the images considered. It was observed that the optimal predictor order and the optimal block size for adaptive linear predictive coding are higher compared to those for the same images digitized to lower spatial resolution. 1.