PoCGM: Poisson-Conditioned Generative Model for Sparse-View CT Reconstruction

2025-11-17

AI Summary

Sparse-view CT reconstruction suffers from severe artifacts and loss of fine anatomical details due to insufficient projection data, limiting its clinical utility in low-dose and dynamic imaging. To address this, we propose the Poisson Conditional Generative Model (PoCGM), the first application of PFGM++ to medical image reconstruction. PoCGM establishes a conditional generative framework wherein sparse projections explicitly guide posterior sampling, incorporating both the Poisson statistics of X-ray photon counting and physical constraints of the imaging geometry. We extend the PFGM++ architecture and integrate physics-informed conditional training and sampling strategies to ensure fidelity to the measurement model. Quantitative evaluation across multiple scales demonstrates that PoCGM significantly outperforms state-of-the-art baselines—achieving superior artifact suppression and higher anatomical structure fidelity—making it particularly suitable for low-dose and time-critical CT applications.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

In computed tomography (CT), reducing the number of projection views is an effective strategy to lower radiation exposure and/or improve temporal resolution. However, this often results in severe aliasing artifacts and loss of structural details in reconstructed images, posing significant challenges for clinical applications. Inspired by the success of the Poisson Flow Generative Model (PFGM++) in natural image generation, we propose a PoCGM (Poisson-Conditioned Generative Model) to address the challenges of sparse-view CT reconstruction. Since PFGM++ was originally designed for unconditional generation, it lacks direct applicability to medical imaging tasks that require integrating conditional inputs. To overcome this limitation, the PoCGM reformulates PFGM++ into a conditional generative framework by incorporating sparse-view data as guidance during both training and sampling phases. By modeling the posterior distribution of full-view reconstructions conditioned on sparse observations, PoCGM effectively suppresses artifacts while preserving fine structural details. Qualitative and quantitative evaluations demonstrate that PoCGM outperforms the baselines, achieving improved artifact suppression, enhanced detail preservation, and reliable performance in dose-sensitive and time-critical imaging scenarios.

Problem

• Reconstructs CT images from sparse projection views
• Reduces radiation exposure while preserving structural details
• Suppresses aliasing artifacts in dose-sensitive imaging scenarios

Innovation

• PoCGM adapts PFGM++ for conditional CT reconstruction
• It integrates sparse-view data during training and sampling
• Models posterior distribution to suppress artifacts and preserve details