Fig 1.
Workflow for developing a deep learning-based autodetection model that detect hyphae in microscopic images obtained from real-world practice.
Table 1.
Summary of fungus hyphae dataset.
Fig 2.
Precision-recall (PR) curves and receiver operating characteristic (ROC) curves with test datasets.
Fig 3.
Example images of the autodetection of hyphae with bounding box.
¥ (A) positive case with 100× magnification, (B) positive case with 40× magnification, (C) negative case with 100× magnification, and (D) negative case (false detection) with 40× magnification. ¥The ground truth were marked with a green box in positive case (A, B).
Fig 4.
Confusion matrix box for (A) Dataset-100, (B) Dataset-40, and (C) all datasets.
Table 2.
Summary of the IOU, TP, FP, FN, precision, recall, F1-score, AP, and AUC values of our model.