Loss odyssey in medical image segmentation

By Jun Ma, Jianan Chen, Matthew Ng, Rui Huang, Yu Li, Chen Li, Xiaoping Yang, Anne L. Martel. Medical Image Anal., 2021

592 citations recorded by the source corpus.

The loss function is an important component in deep learning-based segmentation methods. Over the past five years, many loss functions have been proposed for various segmentation tasks. However, a systematic study of the utility of these loss functions is missing. In this paper, we present a comprehensive review of segmentation loss functions in an organized manner. We also conduct the first large-scale analysis of 20 general loss functions on four typical 3D segmentation tasks involving six public datasets from 10+ medical centers. The results show that none of the losses can consistently achieve the best performance on the four segmentation tasks, but compound loss functions (e.g. Dice with TopK loss, focal loss, Hausdorff distance loss, and boundary loss) are the most robust losses. Our code and segmentation results are publicly available and can serve as a loss function benchmark. We hope this work will also provide insights on new loss function development for the community.

Subfields: Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/0bfa3db4ad2a5f9d1f1926842a0c0e1f4578b3c0

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

AbdomenCT-1K: Is Abdominal Organ Segmentation a Solved Problem? (2020, background): “been proved to be robust in many segmentation tasks [93].”
U-Mamba: Enhancing Long-range Dependency for Biomedical Image Segmentation (2024, methodology): “U-Mamba was also optimized with stochastic gradient descent and the loss function was the unweighted sum of Dice loss and cross-entropy because the compound loss has proven to be robust across different tasks [31].”
Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation (2021, methodology): “State-of-the-art solutions typically use unmodified versions of either the Dice loss, cross entropy loss or a combination of the two, and even when using available loss functions for handling class imbalance, such as the Focal Tversky loss, consistently improved performance has not been observed (Ma”
Segment Anything in Medical Images (2023, background): “The loss function is the unweighted sum between Dice loss and cross-entropy loss, which has been proven to be robust in various segmentation tasks [21][31].”