False claims estimator

Using a Bayesian approach, this tool estimates the probability that a reported outperformance claim is false. It evaluates whether the observed difference between two methods could have occurred by chance, given the test-set size and the reported performance values. The methodology adapts to the type of task: for classification, it models case-wise agreement patterns between methods; for segmentation, it incorporates both performance variability and correlation between methods' per-case scores. The estimator currently supports Accuracy-based comparison for classification and Dice Similarity Coefficient (DSC) values for segmentation.