Segmenting Breast Cancer Pathological Slides

Aim of the project

• Segmenting cells in sections of breast tissue biopsies to help diagnose breast cancer. Specifically a carcinomas under a type called TNBC.
• Aid the pathologist in attaining better accuracies yet while consuming less time and effort.

Description of the Data

• NUCLS datasets :https://nucls.grand-challenge.org/
• We worked with the single rater with 1,744 FOVs | 59,485 nuclei

Model Architecture :