Case ID: M23-189L

Published: 2024-01-02 20:56:30

Last Updated: 1704228990


Wenhui Zhu
Peijie Qiu
Oana Dumitrascu
Yalin Wang

Technology categories

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

Licensing Contacts

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

Framework for Enhancing Retinal Color Fundus Images

Non-mydriatic retinal color fundus photography (CFP) is widely used in clinical settings, primarily because it doesn’t require pupillary dilation. However, CFP has quality issues due to operator errors, imperfections and patient-related causes. Further, medical providers often have difficulty collecting paired low-high quality CFP. Most denoising frameworks are based on supervised learning and require a pair of data (low quality and high quality). Diagnostic interpretations require high-quality CFPs, but non-mydriatic CFPs are prone to noise and artifacts, leading to inaccurate diagnostic interpretations. Enhancing low-quality retinal CFPs into high-quality counterparts is important for many downstream tasks, e.g., blood vessel segmentation, lesion segmentation, diagnostic stratification and more.
Researchers at Arizona State University in collaboration with additional researchers have developed a two-stage framework for retinal CFP enhancement. This framework is based on optimal transport (OT) and regularization by denoising methods. It maximally preserves structural consistency (e.g. lesions, vessel structures, optical discs) between enhanced and low-quality images to prevent over-tampering important structures. This framework eliminates the need for paired data while maximizing the preservation of lesion features in fundus images.
This technology establishes a rigorous measurement pipeline and has broad commercial potential.
Potential Applications
  • Enhancing retinal color fundus photography
    • Blood vessel segmentation
    • Lesion segmentation
    • Diagnostic stratification
    • Disease screening
Benefits and Advantages
  • Solves the problem of hard collecting a pair of high-quality and correspondingly low-quality
  • Preserves maximal lesion information
  • Integrates an OT enhancing network with regularization
  • Has been tested on three datasets with promising results surpassing or on par with existing state-of-the-art techniques
  • Can be used in other eye disease studies to improve image quality and help practitioners aid in their diagnoses
For more information about this opportunity, please see
For more information about the inventor(s) and their research, please see