$mu$NeuFMT: Optical-Property-Adaptive Fluorescence Molecular Tomography via Implicit Neural Representation

2025-11-06

AI Summary

Fluorescence molecular tomography (FMT) reconstruction is severely limited by inaccurate or unknown tissue optical parameters, while existing supervised deep learning methods suffer from poor generalizability. To address this, we propose a self-supervised joint optimization framework that— for the first time—couples implicit neural representations (INRs) with a differentiable photon propagation physics model. This enables simultaneous reconstruction of fluorophore distribution and tissue optical parameters without ground-truth labels or precise prior knowledge of optical properties. Leveraging end-to-end differentiable rendering, our method achieves fully unsupervised optimization, eliminating reliance on large-scale annotated datasets or pre-specified parameters. We validate its robustness across numerical simulations, phantom experiments, and in vivo studies: even with initial optical parameter errors up to 50%, our approach significantly outperforms conventional model-based reconstruction and supervised deep learning methods.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search
• Machine Learning: Unsupervised & Self-Supervised Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Fluorescence Molecular Tomography (FMT) is a promising technique for non-invasive 3D visualization of fluorescent probes, but its reconstruction remains challenging due to the inherent ill-posedness and reliance on inaccurate or often-unknown tissue optical properties. While deep learning methods have shown promise, their supervised nature limits generalization beyond training data. To address these problems, we propose $mu$NeuFMT, a self-supervised FMT reconstruction framework that integrates implicit neural-based scene representation with explicit physical modeling of photon propagation. Its key innovation lies in jointly optimize both the fluorescence distribution and the optical properties ($mu$) during reconstruction, eliminating the need for precise prior knowledge of tissue optics or pre-conditioned training data. We demonstrate that $mu$NeuFMT robustly recovers accurate fluorophore distributions and optical coefficients even with severely erroneous initial values (0.5$ imes$ to 2$ imes$ of ground truth). Extensive numerical, phantom, and in vivo validations show that $mu$NeuFMT outperforms conventional and supervised deep learning approaches across diverse heterogeneous scenarios. Our work establishes a new paradigm for robust and accurate FMT reconstruction, paving the way for more reliable molecular imaging in complex clinically related scenarios, such as fluorescence guided surgery.

Problem

• Reconstructs fluorescence distribution without prior optical properties
• Eliminates need for supervised training data in tomography
• Jointly optimizes fluorescence and optical coefficients robustly

Innovation

• Self-supervised framework integrating neural representation
• Jointly optimizes fluorescence distribution and optical properties
• Eliminates need for precise prior optical knowledge