Initialization and stopping affect accuracy in skin lesion segmentation
Initialization and Stopping Tolerance in CPU Dermoscopic Segmentation
Computer Vision and Pattern Recognition
Summary
This paper looks at how two settings—how you start the outlining process and when you stop it—impact the accuracy of detecting skin lesions in images. The authors used a common tool on a set of skin photos and compared different starting methods and stop rules. They found that starting with a shape based on thresholding the image tends to give better results, but this advantage can change depending on how strictly the process is stopped. It shows that evaluating these methods should always include the original starting outline for comparison and carefully report settings and iteration counts.
What this means in practice
- •For medical image engineers: Optimize skin lesion segmentation pipelines by selecting initial contour and stopping criteria to improve accuracy and reproducibility.
- •For software developers in image analysis: Implement clearer reporting of initialization and stopping parameters to ensure fair comparisons between segmentation methods in tools and apps.
Tested on one dataset.
Authors
Wenhao Xu, Yixian Kong, Ting Pan, Changwei Wang, Feilong Wang, Rongtao Xu
Abstract
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.