Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The authors investigate the interaction between contour initialization (checkerboard, disk, Otsu) and numerical stopping tolerances using scikit-image's Chan-Vese active contour implementation on 600 held-out ISIC 2017 dermoscopy images. Although Otsu initialization significantly improves mean Dice over checkerboard seeds under default tolerance (0.6660 vs. 0.6011), iterative refinement actually degrades performance compared to unrefined Otsu thresholding alone (0.6897), while disk initializers prematurely terminate after a single step in 78.5% of cases. These findings reveal that standard iterative active-contour refinement can mask baseline degradation through numerical stopping artifacts, demanding that vision benchmarks report raw seed baselines, exact tolerances, and actual iteration counts.
Iterative Chan-Vese refinement actually degrades segmentation accuracy compared to raw Otsu thresholding, exposing how uncalibrated stopping tolerances create illusory algorithmic improvements over poorly initialized baselines.
Contour initialization and numerical stopping can jointly affect the evaluation of active-contour segmentation. We examine their interaction using the open-source scikit-image Chan-Vese implementation on a resized ISIC 2017 mirror. A fixed development set of 100 images selects a common input channel; all 600 images in the repository's held-out partition are then evaluated. Otsu thresholding is compared with checkerboard-, disk-, and Otsu-initialized contours under default and tighter level-set tolerances. At the default tolerance, Otsu initialization increases mean image Dice from 0.6011 to 0.6660 relative to checkerboard initialization, a paired difference of 0.0649 (95% image-bootstrap interval [0.0452, 0.0860]). Otsu thresholding alone achieves 0.6897. The default disk initializer stops after one iteration on 471 images. Tightening the tolerance reduces the Otsu-seed advantage over checkerboard initialization to 0.0197, with most runs reaching the 500-iteration limit. The default-tolerance advantage also reverses between small- and large-lesion strata. These findings show that an improvement over a generic initializer can coexist with deterioration relative to the threshold baseline. Evaluations should retain the unrefined mask as a comparator and report the initial-field definition, stopping tolerance, and observed iteration counts together.