On the Optimal Combination of Cross-Entropy and Soft Dice Losses for Lesion Segmentation with Out-of-Distribution Robustness

Date

2023

Authors

Galdran, A.
Carneiro, G.
Ballester, M.A.G.

Editors

Advisors

Journal Title

Journal ISSN

Volume Title

Type:

Conference paper

Citation

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2023, vol.13797 LNCS, pp.40-51

Statement of Responsibility

Adrian Galdran, Gustavo Carneiro and Miguel A. González Ballester

Conference Name

3rd Diabetic Foot Ulcer Challenge (DFUC) (22 Sep 2022 - 22 Sep 2022 : Singapore)

Abstract

We study the impact of different loss functions on lesion segmentation from medical images. Although the Cross-Entropy (CE) loss is the most popular option when dealing with natural images, for biomedical image segmentation the soft Dice loss is often preferred due to its ability to handle imbalanced scenarios. On the other hand, the combination of both functions has also been successfully applied in these types of tasks. A much less studied problem is the generalization ability of all these losses in the presence of Out-of-Distribution (OoD) data. This refers to samples appearing in test time that are drawn from a different distribution than training images. In our case, we train our models on images that always contain lesions, but in test time we also have lesion-free samples. We analyze the impact of the minimization of different loss functions on in-distribution performance, but also its ability to generalize to OoD data, via comprehensive experiments on polyp segmentation from endoscopic images and ulcer segmentation from diabetic feet images. Our findings are surprising: CE-Dice loss combinations that excel in segmenting in-distribution images have a poor performance when dealing with OoD data, which leads us to recommend the adoption of the CE loss for these types of problems, due to its robustness and ability to generalize to OoD samples. Code associated to our experiments can be found at https://github.com/agaldran/lesion losses ood.

School/Discipline

Dissertation Note

Provenance

Description

DFUC 2022 is hosted by MICCAI 2022, the 25th International Conference on Medical Image Computing and Computer Assisted Intervention, September 18-22, 2022.

Access Status

Rights

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023

License

Call number

Persistent link to this record