David Belinsky, Ph.D. Research Proficiency Presentation: 'Nuclear Segmentation in Digital Pathology Images'

Dates: 
Friday, August 20, 2021 - 9:30am to 11:45am
Location: 
Zoom - contact events@cs.stonybrook.edu for Zoom info.
Event Description: 
Abstract:

Deep learning has greatly changed the entire field of computer visionand this is especially true for image segmentation. Image Segmentation is a difficult problem, and the expected outcome can varygreatly from domain to domain, which has led to deep learning tak-ing the forefront.

 
With the advent of many successful and reliable methods forimage segmentation, its applications towards medicine are becoming more prominent. In particular the field of pathology has manyneeds for automated image segmentation. One such problem is nuclear segmentation and classification, which is a subset of image segmentation.
 
In this report, I explore existing approaches of image segmentation. Then I propose an improvement to an existing method of nuclear instance segmentation and classification. I then evaluate these methods quantitatively and show that it obtains State of the Art results.
Computed Event Type: 
Mis