Reassessing the Utility of Topology Based Losses for Image Segmentation

ICLR 2025 Workshop ICBINB Submission33 Authors

Track: long paper (up to 4 pages)

Keywords: Segmentation, Thin-tubular structures, Skeleton Recall Loss

TL;DR: For the task of segmentation in computer vision, topology based loss functions have been proposed. We analyze both empirically and theoretically why they don't work.

Abstract:

Image segmentation is an important and widely performed task in computer vision. Accomplishing effective image segmentation in diverse settings, often requires custom model architectures and loss functions. A set of models that specialize in segmenting thin tubular structures are topology preservation based loss functions. These models often utilize a pixel skeletonization process claimed to generate more precise segmentation masks of thin tubes and better capture the structures other models often missed.

One such model, \ac{SRL} proposed by Kirchhoff et al \citep{kirchhoff2024srl}, was stated to produce state-of-the-art results on benchmark tubular datasets. In this work, we tested the validity of the SRL loss by using two approaches: empirically and theoretically. Upon comparing the performance of the proposed method on some of the tubular datasets (used in the original work, along with some additional datasets), we found that the performance of SRL based segmentation models did not exceed traditional baseline models. We then go on to examine and provide a theoretical explanation as to why losses based on topology based enhancements (including the SRL) fail to fulfill their objective.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Submission Number: 33