top of page

CURVAS: Calibration and Uncertainty for multiRater Volume Assessment in multiorgan Segmentation


Welcome to the home page of the Challenge named CURVAS (Calibration and Uncertainty for multiRater Volume Assessment in multiorgan Segmentation). This challenge is organized by SYCAI TECHNOLOGIES SL in Barcelona (Spain), Universitätsklinikum Erlangen ( Erlangen, Germany) and by the Universitat Pompeu Fabra (UPF) in Barcelona, Spain.This challenge aims to propose a dataset and a source code to calculate uncertainty maps for the segmentation of liver, kidneys and pancreas on CT scan images using Deep Learning (DL) techniques. The participants will have to propose their own source code for the segmentation of the target organs and their lesions using different types of uncertainty maps calculation.


The dataset proposed in this challenge is formed by up to 100 pseudonimized CT scan images belonging to patients with and without lesions in their liver, kidneys and/or pancreas. These lesions are searched to be of three categories depending on their size: small, medium and large. This dataset has been gathered by UKER Hospital (Universitätsklinikum Erlangen). This dataset will be provided to the participants together with theirs labels.

Evaluation and rankingOnce the starting date of the challenge arrives, the dataset will be made public and the participants will have 4 weeks to submit their code to surpass the performance indicators proposed by the organizers to solve the challenge. After these 4 weeks, the organizers will replicate and evaluate the results during the following 3 weeks.After this time, 3 prizes will be awarded: 1000€ to the winner (participant whose code achieves the best results in terms of precission and execution time), 750€ to the second and 300€ to the third participant with best results.


This challenge is expected to open on the 1st. April 2024.


Sycai Technologies SL (Sycai Medical)            Universitätsklinikum Erlangen                     Universitat Pompeu Fabra

bottom of page