 |
This workshop is partially supported
Summary of the Results
There are totally 13 teams registered the challenge this year, and 2 of them successfully submitted algorithms.
Evaluation Method
Dataset: 9 multi-layer cervical cell volumes are used in testing.
Good Segmentation: a ‘‘good’’ segmentation is considered to be one with a DC>0.7.
Abbreviations:
FNo: object-based False negative rate
TPp: pixel-based True positive rate
FPp: pixel-based False positive rate
Quantitative Comparison
Table II - Cytoplasm
| Dice | FNo | TPp | FPp | |
|---|---|---|---|---|
| Phoulady et al. | 0.831±0.079 | 0.408±0.163 | 0.927±0.098 | 0.003±0.002 |
| Ushizima et al. | 0.856±0.078 | 0.501±0.180 | 0.899±0.113 | 0.002±0.001 |
Ranking
Hady Ahmady Phoulady, Dmitry B. Goldgof, Lawrence O. Hall, Peter R. Mouton. AN APPROACH FOR OVERLAPPING CELL SEGMENTATION IN MULTI-LAYER CERVICAL CELL VOLUMES. PDF
Geraldo L. B. Ramalho, Daniel S. Ferreira, Andrea G. C. Bianchi, Claudia M. Carneiro, Fátima N. S. Medeiros, Daniela M. Ushizima. CELL RECONSTRUCTION UNDER VORONOI AND ENCLOSING ELLIPSES FROM 3D MICROSCOPY. PDF