YOLO-UNet Architecture for Detecting and Segmenting the Localized MRI Brain Tumor Image

Table 6

The comparison of average CCR for each methodology.

Methodology	Original image	Gaussian noise	Speckle noise
YOLOv3-UNet	0.978519	0.973193	0.973546
YOLOv4-UNet	0.958206	—	0.936050
UNet model 1	0.970946	0.959026	0.969738
UNet model 2	0.969377	0.959026	0.969678
UNet model 3	0.975070	0.965331	0.976370
UNet model 4	0.973683	0.959354	0.974498
Mask R-CNN	0.964181	0.964679	0.964409

the maximum value of CCR between all methods.