9 December 2024 in IEEE Access
Recently, learning-based lossless compression methods for volumetric medical images have attracted much attention. They can achieve higher compression ratios than traditional methods, albeit at the cost of slower compression speed. Although using field programmable gate arrays (FPGAs) is feasible to mitigate this disadvantage, existing FPGA-based compression frameworks still need CPU for co-processing. In this work, we propose a hardware end-to-end neural conditional entropy encoder (HENCE), for losslessly compressing 3D medical images with the balanced compression ratio and speed. To achieve this, we first introduce a context-based entropy model to reduce data redundancy within inter-slice and intra-slice features, using an efficient combination of auto-regressive and recurrent neural networks. Then, we design a hardware-friendly arithmetic coding module to collaborate with our learning-based entropy model. To obtain the cumulative distribution function of the discrete logistic distribution, we further introduce a high-precision Sigmoid approximation algorithm, using the Newton-Raphson method. Finally, we design a dataflow mechanism for the entropy model and the coding module, achieving a fully pipelined compression system. Extensive experimental results show that our method outperforms traditional image/video codecs like FLIF, JPEX-XL, and HEVC on several volumetric medical image datasets. And our method obtains faster encoding speed than existing learning-based medical image compression frameworks.
8 August 2025 in IEEE Access
In recent years, the continuous advancement of digital technologies such as telemedicine and medical cloud computing has promoted collaborative research and diagnosis across multiple medical centers. However, the timely remote transmission and analysis of large volumetric medical images still pose significant challenges. While classical methods predominantly employing lossless compression are increasingly constrained by the limits of compression ratios, lossy 3D medical image compression methods are emerging as a promising alternative. Different from the existing 3D convolutional compression algorithms oriented only for human vision, this paper proposes a Multi-scale Contextual Autoencoder (MCAE) architecture that recurrently incorporates anatomical inter-slice context to optimize the compression of the current slice for both human and machine vision. Our decoded intermediate features, with sufficiently preserved semantic information, enable high-quality visualization and allow downstream machine vision tasks (e.g., segmentation and classification) to be performed directly without pixel-level recovery. To reduce the compression bit cost, we create a Multi-Dimensional Entropy Model that integrates inter-slice latent context with spatial-channel context and hierarchical hypercontext. Experimental results demonstrate that our framework obtains an average 9% BD-Rate reduction over the Versatile Video Coding (VVC) anchor on MRNet datasets, while achieving superior recognition performance for downstream segmentation and classification tasks than inputting reconstructed lossy images.
1 January 2026 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems