Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 27 Nov 2023]
Title:Streaming Lossless Volumetric Compression of Medical Images Using Gated Recurrent Convolutional Neural Network
View PDFAbstract:Deep learning-based lossless compression methods offer substantial advantages in compressing medical volumetric images. Nevertheless, many learning-based algorithms encounter a trade-off between practicality and compression performance. This paper introduces a hardware-friendly streaming lossless volumetric compression framework, utilizing merely one-thousandth of the model weights compared to other learning-based compression frameworks. We propose a gated recurrent convolutional neural network that combines diverse convolutional structures and fusion gate mechanisms to capture the inter-slice dependencies in volumetric images. Based on such contextual information, we can predict the pixel-by-pixel distribution for entropy coding. Guided by hardware/software co-design principles, we implement the proposed framework on Field Programmable Gate Array to achieve enhanced real-time performance. Extensive experimental results indicate that our method outperforms traditional lossless volumetric compressors and state-of-the-art learning-based lossless compression methods across various medical image benchmarks. Additionally, our method exhibits robust generalization ability and competitive compression speed
Current browse context:
cs
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.