HERALD OF MEDICINE

WANG Yuqun, MENG Mingzhu

Herald of Medicine, 2024, 43(4): 661-667. DOI: 10.3870/j.issn.1004-0781.2024.04.030

Tab.1 Comparison of YOLO-V5 among four sub-models

Other figure/table from this article

Fig.1 A schematic view of the loss function and IoU/GIoU calculation
Fig.2 Brief schematic diagram of YOLO-V5 The model consisted of 2 segments:backbone and head,features of the images were extracted from various dimensions,and the final layers spit out the output.
Tab.2 Hyper parameter settings of YOLO-V5
Fig.3 Flowchart of artificial-intelligence assisted drug dispensation system for pharmacist
Fig.4 Learning curves of YOLO-V5x train/val/box_loss:bounding box detection loss of training set/testing set;train/val/obj_loss:objects detection loss of training set/testing set;train/val/cls_loss:Classification loss of training set/testing set.mAP_{_0.5:0.95}:denotes the mean mAP at the IoU threshold of 0.5~0.95 (from 0.5 to 0.95,step width 0.05,that is 0.5,0.55,0.6,0.65,0.7,0.75,0.8,0.85,0.9,0.95).mAP_{_0.5}:mAP with IoU threshold set at 0.5.
Fig.5 Precision result of YOLO-V5x The blue curve is the average of all classes.
Fig.6 Precision-recall curve of YOLO-V5x The blue curve is the average of all classes. mAP means the area under the curve.mAP_{_0.5}:for details see Fig4.
Tab.3 Comparation of YOLO-V5 among four submodels
Fig.7 Instantiation detection of YOLO-V5x All three drugs were detected correctly.Numbers 45-47 are for the drug classes.0.8 and 0.9 were confidence value.