Enhancing Medical Image Segmentation: Ground Truth Optimization through Evaluating Uncertainty in Expert Annotations

• Published in: Mathematics (Basel) Vol. 11; no. 17; p. 3771
• Main Authors: Athanasiou, Georgios, Arcos, Josep Lluis, Cerquides, Jesus
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.09.2023
• Subjects: Algorithms; Annotations; Artificial neural networks; assisted reproductive technology; convolutional neural networks; Data mining; Datasets; deep learning; Embryonic development; Evaluation; Gametocytes; Image annotation; Image enhancement; Image processing; Image segmentation; Labels; Machine learning; Mathematics; medical data; Medical imaging; Medical imaging equipment; Methods; Neural networks; Optimization; Performance prediction; Reproductive technologies; Supervised learning; Uncertainty
• ISSN: 2227-7390; 2227-7390
• DOI: 10.3390/math11173771

The surge of supervised learning methods for segmentation lately has underscored the critical role of label quality in predicting performance. This issue is prevalent in the domain of medical imaging, where high annotation costs and inter-observer variability pose significant challenges. Acquiring labels commonly involves multiple experts providing their interpretations of the “true” segmentation labels, each influenced by their individual biases. The blind acceptance of these noisy labels as the ground truth restricts the potential effectiveness of segmentation algorithms. Here, we apply coupled convolutional neural network approaches to a small-sized real-world dataset of bovine cumulus oocyte complexes. This is the first time these methods have been applied to a real-world annotation medical dataset, since they were previously tested only on artificially generated labels of medical and non-medical datasets. This dataset is crucial for healthy embryo development. Its application revealed an important challenge: the inability to effectively learn distinct confusion matrices for each expert due to large areas of agreement. In response, we propose a novel method that focuses on areas of high uncertainty. This approach allows us to understand the individual characteristics better, extract their behavior, and use this insight to create a more sophisticated ground truth using maximum likelihood. These findings contribute to the ongoing discussion of leveraging machine learning algorithms for medical image segmentation, particularly in scenarios involving multiple human annotators.