The abstract is read at the publisher; the summary is JClub's.
ResNet50 transfer learning achieved 100% overall accuracy for identifying seven filamentous fungi using culture and grid images from previously unseen strains.
- Why it matters: Culture morphology and microscopy are needed for mould identification, but manual workflows are slow; automated classification requires strong performance across image modalities.
- What they did: The study trained ResNet50 on 15,400 images from seven fungi, using culture-only (7,000), microscopy-only (7,000), ensemble (14,000), and combined culture+microscopy grid (7,000), then tested on unseen strains.
- The result: Culture+grid models reached 100% overall accuracy, while microscopy-only and ensemble approaches reached 99.4% and 98.8%, respectively, enabling accurate image-based fungal identification.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.