Optimization for medical image segmentation: Theory and practice when evaluating with Dice Score or Jaccard Index
| Autor: | Frederik Maes, Jeroen Bertels, Maxim Berman, Raf Bisschops, Matthew B. Blaschko, Tom Eelbode, Dirk Vandermeulen |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Diagnostic Imaging
FOS: Computer and information sciences Computer Science - Machine Learning Jaccard index Computer science Entropy Computer Vision and Pattern Recognition (cs.CV) Computer Science - Computer Vision and Pattern Recognition Dice 030218 nuclear medicine & medical imaging Machine Learning (cs.LG) 03 medical and health sciences 0302 clinical medicine Similarity (network science) FOS: Electrical engineering electronic engineering information engineering Electrical and Electronic Engineering Radiological and Ultrasound Technology business.industry Image and Video Processing (eess.IV) Pattern recognition Image segmentation Electrical Engineering and Systems Science - Image and Video Processing Computer Science Applications Weighting Metric (mathematics) Artificial intelligence business Software |
| Popis: | In many medical imaging and classical computer vision tasks, the Dice score and Jaccard index are used to evaluate the segmentation performance. Despite the existence and great empirical success of metric-sensitive losses, i.e. relaxations of these metrics such as soft Dice, soft Jaccard and Lovasz-Softmax, many researchers still use per-pixel losses, such as (weighted) cross-entropy to train CNNs for segmentation. Therefore, the target metric is in many cases not directly optimized. We investigate from a theoretical perspective, the relation within the group of metric-sensitive loss functions and question the existence of an optimal weighting scheme for weighted cross-entropy to optimize the Dice score and Jaccard index at test time. We find that the Dice score and Jaccard index approximate each other relatively and absolutely, but we find no such approximation for a weighted Hamming similarity. For the Tversky loss, the approximation gets monotonically worse when deviating from the trivial weight setting where soft Tversky equals soft Dice. We verify these results empirically in an extensive validation on six medical segmentation tasks and can confirm that metric-sensitive losses are superior to cross-entropy based loss functions in case of evaluation with Dice Score or Jaccard Index. This further holds in a multi-class setting, and across different object sizes and foreground/background ratios. These results encourage a wider adoption of metric-sensitive loss functions for medical segmentation tasks where the performance measure of interest is the Dice score or Jaccard index. 15 pages, 14 figures, accepted for publication in IEEE Transactions on Medical Imaging (2020) |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|