Skip to content

Latest commit

 

History

History
29 lines (23 loc) · 1.05 KB

README.md

File metadata and controls

29 lines (23 loc) · 1.05 KB

Sementic-Segmentation on LIVECell Dataset

LIVECell Dataset (https://sartorius-research.github.io/LIVECell/)

LIVECell is a large, high-quality, manually annotated and expert-validated dataset of phase-contrast images, consisting of over 1.6 million cells from a diverse set of cell morphologies and culture densities.

Model used : CANet (Chained Context Aggregation Network) with few modifications.

  • Loss Used : dice_loss + (1 * focal_loss)
  • Total Epochs : 10
  • Training IOU Score : 0.6533
  • Validation IOU Score : 0.7483
  • Training F1 Score : 0.7812
  • Validation F1 Score : 0.8532

Model Performance.

Visualization of model prediction.

Reference :

  1. https://www.nature.com/articles/s41592-021-01249-6
  2. https://www.sciencedirect.com/science/article/abs/pii/S0262885621002146