Histological techniques are valuable tools in various biological and medical disciplines. They have proven particularly valuable in ecotoxicological studies, providing critical insight into the sublethal effects of pollutants on aquatic organisms. However, limited reference data on invertebrate histology make these analyses time-consuming, particularly for previously undescribed species. In this study, a deep learning model, based on the Convolutional Neural Network (CNN) was developed to automatically identify 11 tissue types of Chironomus riparius. The model achieved an overall accuracy of 94.21%, with five of the eleven tissues correctly identified in all cases, while the highest misclassification rate (22.72%) occurred in the case of recognizing parietal fat body. This represents the first AI-assisted approach for histological tissue classification in a standard invertebrate Organisation for Economic Co-operation and Development model organism, achieving high accuracy and providing a foundation for integrating deep learning into future histopathological workflows in environmental and ecotoxicological research.
Stojanović et al. (Wed,) studied this question.