Case ID: M19-117L^

Published: 2020-02-26 10:43:32

Last Updated: 1677135956


Inventor(s)

Md Mahfuzur Rahman Siddiquee
Zongwei Zhou
Ruibin Feng
Nima Tajbakhsh
Jianming Liang

Technology categories

Computing & Information TechnologyImagingLife Science (All LS Techs)Medical Imaging

Licensing Contacts

Jovan Heusser
Director of Licensing and Business Development
[email protected]

Fixed-Point Generative Adversarial Networks

Generative adversarial networks (GANs) are revolutionizing image-to-image translation, which is attractive to researchers in the medical imaging community.  While using GANs to reveal diseased regions in a medical image is appealing, it requires a GAN to identify a minimal subset of target pixels for domain translation, also known as fixed-point translation, which is not possible with current GANs.

 

Researchers at Arizona State University have proposed a new GAN, called Fixed-Point GAN, which introduces fixed-point translation and proposes a new method for disease detection and localization. This new Gan is trained by (1) supervising same-domain translation through a conditional identity loss, and (2) regularizing cross-domain translation through revised adversarial, domain classification, and cycle consistency loss. Qualitative and quantitative evaluations demonstrate that the proposed method outperforms the state of the art in multi-domain image-to-image translation and that it surpasses predominant weakly-supervised localization methods in both disease detection and localization.

 

This fixed-point GAN dramatically reduces artifacts in image-to-image translation and introduces a novel method for disease detection and localization that outperforms the state of the art.

 

Potential Applications

•       Computer-aided diagnoses – e.g. pulmonary embolism and brain lesion localization

•       Non-medical applications – photo editing/aging/blending, game development and animation production

 

Benefits and Advantages

•       Works with unpaired images – does not require two images with and without the attribute

•       Requires only image-level annotation for training

•       Same-domain translation without adding or removing attributes

•       Cross-domain translation without affecting unrelated attributes

o       E.g. removes eyeglasses from an image without affecting hair color

•       Source-domain-independent translation using only image-level annotation

•       Outperforms the state of the art in multi-domain image-to-image translation for both natural and medical images

•       Surpasses predominant weakly-supervised localization methods in both disease detection and localization

•       Dramatically reduces artifacts in image-to-image translation

 

For more information about this opportunity, please see

Rahman Siddiquee et al – ICCV – 2019

GitHub – 2019

For more information about the inventor(s) and their research, please see

Dr. Liang’s departmental webpage