Spectral Bias Correction in PINNs for Myocardial Image Registration of Pathological Data

2025-04-24

AI Summary

Physics-informed neural networks (PINNs) suffer from spectral bias, hindering accurate modeling of high-frequency deformations in pathological cardiac image registration—leading to registration inaccuracies and biomechanically implausible deformations. Method: This work proposes the first integration of Fourier feature mapping with a learnable modulation mechanism into the PINN framework to explicitly mitigate spectral bias and enhance representation of pathological high-frequency deformations. The method combines Fourier feature encoding, a modulated PINN architecture, biomechanics-aware loss functions, and physics-based deformation modeling. Results: Evaluated on two pathological cardiac datasets, the method achieves a 3.2% improvement in Dice similarity coefficient and a 28% reduction in target registration error (TRE). Registered deformations adhere to myocardial biomechanical constraints and demonstrate strong generalization across patients and multiple cardiac pathologies.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

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

Abstract

Accurate myocardial image registration is essential for cardiac strain analysis and disease diagnosis. However, spectral bias in neural networks impedes modeling high-frequency deformations, producing inaccurate, biomechanically implausible results, particularly in pathological data. This paper addresses spectral bias in physics-informed neural networks (PINNs) by integrating Fourier Feature mappings and introducing modulation strategies into a PINN framework. Experiments on two distinct datasets demonstrate that the proposed methods enhance the PINN's ability to capture complex, high-frequency deformations in cardiomyopathies, achieving superior registration accuracy while maintaining biomechanical plausibility - thus providing a foundation for scalable cardiac image registration and generalization across multiple patients and pathologies.

Problem

• Corrects spectral bias in PINNs for myocardial image registration
• Enhances high-frequency deformation modeling in pathological data
• Improves registration accuracy and biomechanical plausibility

Innovation

• Integrates Fourier Feature mappings in PINNs
• Introduces modulation strategies for high-frequency deformations
• Enhances registration accuracy with biomechanical plausibility