An Empirical Study on MC Dropout--Based Uncertainty--Error Correlation in 2D Brain Tumor Segmentation
Saumya B
TL;DR
This study interrogates whether MC Dropout uncertainty can identify segmentation errors in 2D brain tumor MRI and whether data augmentation alters this relationship. Using a vanilla U-Net trained under four augmentations, and 50 MC forward passes per image, the authors quantify both global and boundary uncertainty and correlate them with pixel-wise errors. They find weak global correlations ($r \approx 0.30$–$0.38$) and negligible boundary correlations ($|r| < 0.05$), with augmentation effects statistically significant but practically minor. The results imply that MC Dropout provides limited guidance for boundary error localization, underscoring the need for alternative or hybrid uncertainty estimation approaches in medical image segmentation.
Abstract
Accurate brain tumor segmentation from MRI is vital for diagnosis and treatment planning. Although Monte Carlo (MC) Dropout is widely used to estimate model uncertainty, its effectiveness in identifying segmentation errors -- especially near tumor boundaries -- remains unclear. This study empirically examines the relationship between MC Dropout--based uncertainty and segmentation error in 2D brain tumor MRI segmentation using a U-Net trained under four augmentation settings: none, horizontal flip, rotation, and scaling. Uncertainty was computed from 50 stochastic forward passes and correlated with pixel-wise errors using Pearson and Spearman coefficients. Results show weak global correlations ($r \approx 0.30$--$0.38$) and negligible boundary correlations ($|r| < 0.05$). Although differences across augmentations were statistically significant ($p < 0.001$), they lacked practical relevance. These findings suggest that MC Dropout uncertainty provides limited cues for boundary error localization, underscoring the need for alternative or hybrid uncertainty estimation methods in medical image segmentation.