Home Publications Research Grants Inventions & Patents Awards Additional Research Activities Faculties & Institutions Research Areas

An asymmetric cycle-consistency loss for dealing with many-to-one mappings in image translation: A study on thigh mr scans

Gadermayr M, Tschuchnig M, Gupta L, Kraemer N, Truhn D, Merhof D, Gess B (2021)

---

Publication Type: Conference contribution

Publication year: 2021

Journal

Publisher: IEEE Computer Society

Book Volume: 2021-April

Pages Range: 1182-1186

Conference Proceedings Title: Proceedings - International Symposium on Biomedical Imaging

Event location: Nice, FRA

ISBN: 9781665412469

DOI: 10.1109/ISBI48211.2021.9433891

Abstract

Adversarial networks using a cycle-consistency loss facilitate unpaired training of image-translation models and thereby exhibit a high potential in medical applications. However, the fact that images in one domain potentially map to more than one image in another domain (e.g. in case of pathological changes) exhibits a major challenge for training the networks. We offer a solution to improve the training process in case of many-to-one mappings by modifying the cycle-consistency loss. We show formally and empirically that the proposed method improves the performance without radically changing the architecture and increasing the model complexity.

Involved external institutions

How to cite

APA:

Gadermayr, M., Tschuchnig, M., Gupta, L., Kraemer, N., Truhn, D., Merhof, D., & Gess, B. (2021). An asymmetric cycle-consistency loss for dealing with many-to-one mappings in image translation: A study on thigh mr scans. In Proceedings - International Symposium on Biomedical Imaging (pp. 1182-1186). Nice, FRA: IEEE Computer Society.

MLA:

Gadermayr, M., et al. "An asymmetric cycle-consistency loss for dealing with many-to-one mappings in image translation: A study on thigh mr scans." Proceedings of the 18th IEEE International Symposium on Biomedical Imaging, ISBI 2021, Nice, FRA IEEE Computer Society, 2021. 1182-1186.

BibTeX: Download