Research Article

SK-FMYOLOV3: A Novel Detection Method for Urine Test Strips

Table 4

Performance comparison of different detection models.

ModelImproved prediction boxMultiscaleSkNet (%) (%)Speed/ms

R-CNN [21]78.564.21200
FAST-RCNN [23]86.679.7700
FASTER-RCNN [24]89.887.3350
YOLOV3 [32]87.671.4230
SE-YOLOV392.586.1350
FYOLOV393.190.2310
MYOLOV392.389.6340
FMYOLOV394.691.7360
SK-FMYOLOV396.894.5390