3D Field of Junctions: A Noise-Robust, Training-Free Structural Prior for Volumetric Inverse Problems
2026-03-02 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed a new way to clean up noisy 3D images called 3D Field of Junctions (3D FoJ). It breaks down 3D images into small wedge-shaped pieces and finds the best fit for each patch, keeping these fits consistent across the whole volume. Their method does not need any training data, avoids creating fake details, and helps keep edges and corners sharp even when the images are very noisy. They tested 3D FoJ on different tough 3D imaging tasks like low-dose CT scans, cryo-electron tomography, and noisy lidar data, and found it works better than many existing methods.
Volume denoising3D imagingField of JunctionsInverse problemsLow signal-to-noise ratioX-ray computed tomographyCryogenic electron tomographyPoint cloudsProximal gradient descentStructural prior
Authors
Namhoon Kim, Narges Moeini, Justin Romberg, Sara Fridovich-Keil
Abstract
Volume denoising is a foundational problem in computational imaging, as many 3D imaging inverse problems face high levels of measurement noise. Inspired by the strong 2D image denoising properties of Field of Junctions (ICCV 2021), we propose a novel, fully volumetric 3D Field of Junctions (3D FoJ) representation that optimizes a junction of 3D wedges that best explain each 3D patch of a full volume, while encouraging consistency between overlapping patches. In addition to direct volume denoising, we leverage our 3D FoJ representation as a structural prior that: (i) requires no training data, and thus precludes the risk of hallucination, (ii) preserves and enhances sharp edge and corner structures in 3D, even under low signal to noise ratio (SNR), and (iii) can be used as a drop-in denoising representation via projected or proximal gradient descent for any volumetric inverse problem with low SNR. We demonstrate successful volume reconstruction and denoising with 3D FoJ across three diverse 3D imaging tasks with low-SNR measurements: low-dose X-ray computed tomography (CT), cryogenic electron tomography (cryo-ET), and denoising point clouds such as those from lidar in adverse weather. Across these challenging low-SNR volumetric imaging problems, 3D FoJ outperforms a mixture of classical and neural methods.