Target detection results (the detection results of proposed algorithm are on the left and YOLOv7 is on the middle YOLOv8-n is on the right).

<p>Target detection results (the detection results of proposed algorithm are on the left and YOLOv7 is on the middle YOLOv8-n is on the right).</p>

Saved in:

Bibliographic Details
Main Author:	Xiangqian Xu (17310895) (author)
Other Authors:	Wenting Hou (1995478) (author), Xing Li (134367) (author)
Published:	2024
Subjects:	Medicine Sociology Space Science Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified loss function ciou less computational overhead href =" mailto experimental results show eiou loss function depthwise separable convolution 3 %, respectively feature fusion network div >< p cause missed detection small size defects small defect size model calculation amount improved yolov7 algorithm defect size network calculation parameter amount improved yolov7 convolution module 2 %, backbone network detect small defect images defect detection weld image sized targets reach 72 parameter quantity model volume high degree false detection effectively reduce effectively applied detection method detection ability convergence speed completely detect
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items