Learning Lossless Compression for High Bit-Depth Medical Imaging

We propose a learned lossless image compression method for high bit-depth medical imaging (up to 16 bit-depths). Instead of compressing a high bit-depth medical image as a whole, we split it into two low bit-depth subimages, i.e., the most significant bytes (MSB) subimage and the least significant bytes (LSB) subimage, respectively. The MSB subimage depicts piece-wise smooth structure information that is relatively easy to compress. We thus use traditional lossless codecs for low complexity. The LSB subimage depicts the complementary texture information that is more challenging to compress. We design an autoregressive entropy model conditioned on the MSB subimage that models the probability distribution of the LSB subimage and effectively reduces the redundancy between the MSB and LSB subimages. We then encode the LSB subimage to bitstreams based on the learned entropy model. The compressed high bit-depth medical image is finally stored including the bitstreams of the MSB and LSB subimages. Experimental results demonstrate the state-of-the-art compression performance of the proposed method on high bit-depth medical images, compared with both existing traditional and learned lossless image codecs.

I. Introduction

Medical imaging is an indispensable tool in clinical diagnosis. Nowadays, the usage of digital medical imaging is rising sharply, which leads to an increasingly heavy burden on medical image storage, transmission, and management. Lossy compression used to reduce the bit rates of natural images is unsuitable for medical images, because the artifacts introduced by lossy compression may mislead diagnosis and result in potential medical malpractices. In order to meet the stringent demands on image fidelity, lossless compression is the most reliable choice for medical images, which can perfectly reconstruct original image data from compressed bitstreams. For Computed Tomography (CT) and Magnetic Resonance Imaging (MRI), the most widely used medical imaging format is Digital Imaging and Communications in Medicine (DICOM), which supports traditional lossless image codecs, such as JPEG-LS [1] and JPEG 2000 [2]. These traditional lossless image codecs are originally designed for natural images, of which the compression performance is rather limited.