Panoptic segmentation with highly imbalanced semantic labels

Item Type:	Preprint
Title:	Panoptic segmentation with highly imbalanced semantic labels
Creators Name:	Rumberger, J.L. and Baumann, E. and Hirsch, P. and Janowczyk, A. and Zlobec, I. and Kainmueller, D.
Abstract:	This manuscript describes the panoptic segmentation method we devised for our submission to the CONIC challenge at ISBI 2022. Key features of our method are a weighted loss that we specifically engineered for semantic segmentation of highly imbalanced cell types, and an existing state-of-the art nuclei instance segmentation model, which we combine in a Hovernet-like architecture.
Keywords:	Nuclei Segmentation, Cell Classification, Digital Pathology, Challenge Submission
Source:	arXiv
Publisher:	Cornell University
Article Number:	2203.11692
Date:	19 April 2022
Official Publication:	https://doi.org/10.48550/arXiv.2203.11692

Repository Staff Only: item control page