Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout

2025-01-20

AI Summary

To address biased uncertainty estimation in medical image semantic segmentation, this paper proposes Monte Carlo Frequency-Domain Dropout (MC-FD). During inference, feature maps are transformed via Fast Fourier Transform (FFT), followed by stochastic attenuation of high-frequency components to enhance texture diversity; inverse FFT (iFFT) then reconstructs the perturbed features, preserving structural consistency while improving robustness to input perturbations. MC-FD is the first method to extend stochastic regularization to the frequency domain and integrates Monte Carlo sampling for efficient uncertainty quantification—specifically entropy estimation and confidence calibration. Evaluated on MRI-based prostate segmentation, CT-based liver tumor segmentation, and X-ray lung segmentation, MC-FD significantly improves uncertainty calibration (reducing Expected Calibration Error by 23.6%), accelerates model convergence (by 18.4%), and enhances boundary accuracy (reducing Hausdorff Distance at 95th percentile by 15.2%). These advances collectively strengthen clinical decision-making reliability.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Segmentation
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Monte-Carlo (MC) Dropout provides a practical solution for estimating predictive distributions in deterministic neural networks. Traditional dropout, applied within the signal space, may fail to account for frequency-related noise common in medical imaging, leading to biased predictive estimates. A novel approach extends Dropout to the frequency domain, allowing stochastic attenuation of signal frequencies during inference. This creates diverse global textural variations in feature maps while preserving structural integrity -- a factor we hypothesize and empirically show is contributing to accurately estimating uncertainties in semantic segmentation. We evaluated traditional MC-Dropout and the MC-frequency Dropout in three segmentation tasks involving different imaging modalities: (i) prostate zones in biparametric MRI, (ii) liver tumors in contrast-enhanced CT, and (iii) lungs in chest X-ray scans. Our results show that MC-Frequency Dropout improves calibration, convergence, and semantic uncertainty, thereby improving prediction scrutiny, boundary delineation, and has the potential to enhance medical decision-making.

Problem

• Semantic Segmentation
• Uncertainty Quantification
• Medical Imaging

Innovation

• Monte Carlo Dropout
• Semantic Segmentation
• Uncertainty Estimation