Does the Skeleton-Recall Loss Really Work?

TL;DR

This paper critically evaluates the Skeleton Recall Loss (SRL) for image segmentation of tubular structures, revealing that it does not outperform traditional methods despite claims of state-of-the-art results, through theoretical and empirical analysis.

Contribution

The paper provides a theoretical analysis of SRL's gradients and empirically compares its performance, challenging claims of its superiority in tubular structure segmentation.

Findings

SRL does not outperform baseline models on tubular datasets

Theoretical analysis explains limitations of SRL gradients

Empirical results contradict original claims of state-of-the-art performance

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 that other models often miss. One such model, Skeleton Recall Loss (SRL) proposed by Kirchhoff et al.~\cite {kirchhoff2024srl}, was stated to produce state-of-the-art results on benchmark tubular datasets. In this work, we performed a theoretical analysis of the gradients for the SRL loss. 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. By providing both a theoretical explanation and empirical evidence, this work critically evaluates the limitations of topology-based loss functions, offering valuable insights for researchers aiming to develop more effective segmentation models for complex tubular structures.