J Imaging Inform Med. 2026 Sep 8. doi: 10.1007/s10278-026-02115-7. Online ahead of print.
ABSTRACT
Parasitic helminth infections remain a major public health concern, particularly in low-resource settings where microscopic diagnosis requires significant time and expertise. This study presents ParasiteNet, a deep learning-based framework for automated classification of 12 helminth egg categories from microscopic images. A dataset of 12,000 annotated images was used to train and evaluate two architectures: CNN+ResNet50 hybrid model and EfficientNetV2B0. Image preprocessing and augmentation techniques were applied to improve model robustness, and performance was evaluated using accuracy, precision, recall, F1-score, and loss metrics. Experimental results showed that EfficientNetV2B0 achieved 96.92% classification accuracy with a loss of 0.0067, outperforming the CNN+ResNet50 hybrid model, which achieved 95.25% accuracy with a loss of 0.1840. The proposed framework demonstrates promising performance for automated parasite egg classification under controlled experimental conditions. However, additional validation using independent external datasets, diverse microscopy settings, and prospective clinical studies is necessary to establish the robustness and generalizability of the approach before real-world deployment.
PMID:42711646 | DOI:10.1007/s10278-026-02115-7
