Optical Instruments, Volume. 43, Issue 6, 26(2021)
Single-frame image eyeball tracking based on deformable convolution
Figures & Tables(5)
Fig. 3. Schematic diagram of deformable convolutional YOLO network model
Table 1. Comparison of deformable YOLO network and other YOLO networks
View tableView in ArticleTable 1. Comparison of deformable YOLO network and other YOLO networks
方法 网络框架 精准度 刷新率/（帧/s） YOLO V1 ResNet-101 0.634 45 YOLO V3 ResNet-152 0.579 51 Fast-YOLO Vgg16 0.527 155 可变形卷积YOLO Vgg16 0.685 42
Table 2. Comparison of deformable YOLO network and other advanced networks
View tableView in ArticleTable 2. Comparison of deformable YOLO network and other advanced networks
方法 网络框架 精准度 刷新率/（帧/s） 100 Hz DPM SVM 0.160 100 30 Hz DPM SVM 0.261 30 Fastest DPM SVM 0.304 15 Fast R-CNN Vgg-16 0.701 0.5 R-CNN Minus R Vgg-32 0.535 6 YOLO Vgg-16 Vgg-16 0.654 18 可变形卷积 YOLO Vgg-16 0.685 42
Tools
Get Citation
Copy Citation Text
Jian WANG, Rongfu ZHANG. Single-frame image eyeball tracking based on deformable convolution[J]. Optical Instruments, 2021, 43(6): 26
Paper Information
Category: APPLICATION TECHNOLOGY
Received: Mar. 12, 2021
Accepted: --
Published Online: Jun. 29, 2022
The Author Email: ZHANG Rongfu (zrf@usst.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20