Ultralow‐parameter denoising: Trainable bilateral filter layers in computed tomography
Authors:
Open Access
- 30 May 2022
- journal article
- research article
- Published by Wiley in Medical Physics
- Vol. 49 (8) , 5107-5120
Abstract
Background: Computed tomography (CT) is widely used as an imaging tool to visualize three‐dimensional structures with expressive bone‐soft tissue contrast. However, CT resolution can be severely degraded through low‐dose acquisitions, highlighting the importance of effective denoising algorithms.Purpose: Most data‐driven denoising techniques are based on deep neural networks, and therefore, contain hundreds of thousands of trainable parameters, making them incomprehensible and prone to prediction failures. Developing understandable and robust denoising algorithms achieving state‐of‐the‐art performance helps to minimize radiation dose while maintaining data integrity.Methods: This work presents an open‐source CT denoising framework based on the idea of bilateral filtering. We propose a bilateral filter that can be incorporated into any deep learning pipeline and optimized in a purely data‐driven way by calculating the gradient flow toward its hyperparameters and its input. Denoising in pure image‐to‐image pipelines and across different domains such as raw detector data and reconstructed volume, using a differentiable backprojection layer, is demonstrated. In contrast to other models, our bilateral filter layer consists of only four trainable parameters and constrains the applied operation to follow the traditional bilateral filter algorithm by design.Results: Although only using three spatial parameters and one intensity range parameter per filter layer, the proposed denoising pipelines can compete with deep state‐of‐the‐art denoising architectures with several hundred thousand parameters. Competitive denoising performance is achieved on x‐ray microscope bone data and the 2016 Low Dose CT Grand Challenge data set. We report structural similarity index measures of 0.7094 and 0.9674 and peak signal‐to‐noise ratio values of 33.17 and 43.07 on the respective data sets.Conclusions: Due to the extremely low number of trainable parameters with well‐defined effect, prediction reliance and data integrity is guaranteed at any time in the proposed pipelines, in contrast to most other deep learning‐based denoising architectures.Keywords
All Related Versions
Funding Information
- European Research Council (810316)
This publication has 44 references indexed in Scilit:
- Radiation Exposure from CT Scans: How to Close Our Knowledge Gaps, Monitor and Safeguard Exposure—Proceedings and Recommendations of the Radiation Dose Summit, Sponsored by NIBIB, February 24–25, 2011Radiology, 2012
- Iterative reconstruction methods in X-ray CTPhysica Medica, 2012
- Development and Validation of a Practical Lower-Dose-Simulation Tool for Optimizing Computed Tomography Scan ProtocolsJournal of Computer Assisted Tomography, 2012
- Three‐dimensional anisotropic adaptive filtering of projection data for noise reduction in cone beam CTMedical Physics, 2011
- Bimodal distribution of osteocyte lacunar size in the human femoral cortex as revealed by micro-CTBone, 2010
- Introduction to Bone Development, Remodelling and Repair2008
- Approximation and estimation bounds for artificial neural networksMachine Learning, 1994
- Practical cone-beam algorithmJournal of the Optical Society of America A, 1984
- Statistical limitations in transaxial tomographyComputers in Biology and Medicine, 1976
- Iterative methods for the three-dimensional reconstruction of an object from projectionsJournal of Theoretical Biology, 1972