Progressive and Multipath Holistically Nested Neural Network for Improved Detail Level in Medical Image Segmentation


The technology relates to software, systems, and methods for automated medical image segmentation via deep learning. Standard holistically-nested networks (HNNs) used in computer vision and medical imaging for segmentation and edge detection have a problem with coarsening resolution. The described technology addresses this.

Simple, progressive multipath connections enhance the standard HNN model. Processing in this new model achieves improved segmentations detail levels without need for additional model parameters. Variations in appearance are handled without being affected by variations in shape for a given image feature. Early experiments improved segmentation masks compared to standard HNN. Anatomical regions, often misidentified, are correctly captured via this approach.

This technology is generalizable to image segmentation across multiple imaging modalities, including CT, MRI and X-ray. The improved image segmentation detail levels may have clinically relevant applications to high variability image features, such as tumors and pathologies of the lung. This technology has potential to improve diagnostic capabilities and treatment outcomes in many disease conditions.

Potential Commercial Applications: Competitive Advantages:
  • Computer assisted diagnostics and medical imaging processing
  • Tumor imaging/diagnosis
  • Imaging pathologies of the lung
  • Diagnostic capabilities for a range of disease conditions, e.g., cardiovascular and neurodegenerative
  • Providing framework to develop more cost-effective health care
 
  • Improved detail in automated medical image segmentation without need for additional training or processing of images on other parameters
  • Generalizable to image segmentation across multiple imaging modalities, including CT, MRI and X-ray
  • Improved segmentations detail levels without need for additional model parameters
  • Improved segmentation masks compared to standard HNN


Inventors:

Adam Harrison (CC)  ➽ more inventions...

Ziyue Xu (CC)  ➽ more inventions...

Le Lu (CC)  ➽ more inventions...

Ronald Summers (CC)  ➽ more inventions...

Daniel Mollura (CC)  ➽ more inventions...


Intellectual Property:
US Application No. 62/516,948

Collaboration Opportunity:

Researchers at the NIH Clinical Center seek licensing and development partnership for this technology. Please contact John D. Hewes, Ph.D. at 240-276-5515 or john.hewes@nih.gov for more information.


Licensing Contact:
Edward (Tedd) Fenn, J.D.
Email: tedd.fenn@nih.gov
Phone: 240-276-6833

OTT Reference No: E-109-2017/0
Updated: Nov 17, 2017