Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 26 Oct 2022 (v1), last revised 18 Sep 2024 (this version, v3)]
Title:Residual Back Projection With Untrained Neural Networks
View PDF HTML (experimental)Abstract:Background and Objective: The success of neural networks in a number of image processing tasks has motivated their application in image reconstruction problems in computed tomography (CT). While progress has been made in this area, the lack of stability and theoretical guarantees for accuracy, together with the scarcity of high-quality training data for specific imaging domains pose challenges for many CT applications. In this paper, we present a framework for iterative reconstruction (IR) in CT that leverages the hierarchical structure of neural networks, without the need for training. Our framework incorporates this structural information as a deep image prior (DIP), and uses a novel residual back projection (RBP) connection that forms the basis for our iterations.
Methods: We propose using an untrained U-net in conjunction with a novel residual back projection to minimize an objective function and achieve high-accuracy reconstruction. In each iteration, the weights of the untrained U-net are optimized, and the output of the U-net in the current iteration is used to update the input of the U-net in the next iteration through the aforementioned RBP connection.
Results: Experimental results demonstrate that the RBP-DIP framework offers improvements over other state-of-the-art conventional IR methods, as well as pre-trained and untrained models with similar network structures under multiple conditions. These improvements are particularly significant in the few-view, limited-angle, and low-dose imaging configurations.
Conclusions: Applying to both parallel and fan beam X-ray imaging, our framework shows significant improvement under multiple conditions. Furthermore, the proposed framework requires no training data and can be adjusted on-demand to adapt to different conditions (e.g. noise level, geometry, and imaged object).
Submission history
From: Ziyu Shu [view email][v1] Wed, 26 Oct 2022 01:58:09 UTC (7,976 KB)
[v2] Fri, 29 Dec 2023 21:24:49 UTC (6,514 KB)
[v3] Wed, 18 Sep 2024 00:39:53 UTC (6,600 KB)
Current browse context:
eess.IV
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
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.