🤖 AI Summary
This study addresses the misalignment between instance detection and voxel-level metrics, as well as the frequent miss-detection of small lesions in stroke segmentation. We propose a Volume-Conditioned Adaptive Post-processing (VCAP) scheme alongside the Viola2Plus architecture. By employing dynamic threshold adjustment to bridge detection gaps and integrating resolution-aware attention with a dual-architecture ensemble, our method significantly enhances small target detectability. Notably, results demonstrate that optimized post-processing contributes more to instance detection performance than architectural modifications alone. Five-fold cross-validation achieved a Dice score of 0.651 and a Lesion-F1 of 0.614, with a 3.7% improvement in small lesion detection rate, effectively mitigating the limitations of conventional evaluation metrics.
📝 Abstract
Instance-level lesion detection has been an increasingly larger focal point in medical image segmentation besides the more standard voxel-level overlap. Still, most pipelines are trained and post-processed for voxel overlap alone. In particular, the mismatch is most pronounced for small lesions, where a near-miss prediction---substantial overlap that falls just short of the instance-matching threshold---scores the same as a complete miss. In our ISLES'26 submission, we found that closing this gap mattered far more in post-processing than in architecture design. Our Volume-Conditioned Adaptive Post-Processing (VCAP) scheme adjusts component-size thresholds to each case's predicted lesion burden, improving Lesion-F1 by 0.032 (unbiased cross-fold estimate)---approximately 6 times larger than any architectural change we tested. A resolution-aware attention architecture (Viola2Plus), designed for small-lesion segmentation, shows why the distinction matters: it left small-lesion Dice unchanged but raised small-lesion detection rate by 3.7\%, a real effect voxel-overlap metrics alone would have missed. Under 5-fold cross-validation on the 1,453-case training set, our post-processed two-architecture ensemble achieves Dice 0.651 and Lesion-F1 0.614, versus 0.644 and 0.573 for the unprocessed single-model baseline.