Author
Listed:
- Genjian Yang
(School of Automation, Beijing Information Science and Technology University, Beijing 100192, China)
- Wenbai Chen
(School of Automation, Beijing Information Science and Technology University, Beijing 100192, China)
- Peiliang Wu
(School of Information Science and Engineering, Yanshan University, Qinhuangdao 066004, China)
- Jianping Gou
(College of Computer and Information Science, College of Software, Southwest University, Chongqing 400715, China)
- Xintong Meng
(School of Automation, Beijing Information Science and Technology University, Beijing 100192, China)
Abstract
The precise and robust localization of pupils is crucial for advancing medical diagnostics and enhancing user experience. Currently, the predominant method for determining the center of the pupil relies on the principles of multi-view geometry, necessitating the simultaneous operation of multiple sensors at different angles. This study introduces a single-stage pupil localization network named ResDenseDilateNet, which is aimed at utilizing a single sensor for pupil localization and ensuring accuracy and stability across various application environments. Our network utilizes near-infrared (NIR) imaging to ensure high-quality image output, meeting the demands of most current applications. A unique technical highlight is the seamless integration of the efficient characteristics of the Deep Residual Network (ResNet) with the Dense Dilated Convolutions Merging Module (DDCM), which substantially enhances the network’s performance in precisely capturing pupil features, providing a deep and accurate understanding and extraction of pupil details. This innovative combination strategy greatly improves the system’s ability to handle the complexity and subtleties of pupil detection, as well as its adaptability to dynamic pupil changes and environmental factors. Furthermore, we have proposed an innovative loss function, the Contour Centering Loss, which is specifically designed for irregular or partially occluded pupil scenarios. This method innovatively calculates the pupil center point, significantly enhancing the accuracy of pupil localization and robustness of the model in dealing with varied pupil morphologies and partial occlusions. The technology presented in this study not only significantly improves the precision of pupil localization but also exhibits exceptional adaptability and robustness in dealing with complex scenarios, diverse pupil shapes, and occlusions, laying a solid foundation for the future development and application of pupil localization technology.
Suggested Citation
Genjian Yang & Wenbai Chen & Peiliang Wu & Jianping Gou & Xintong Meng, 2024.
"An Irregular Pupil Localization Network Driven by ResNet Architecture,"
Mathematics, MDPI, vol. 12(17), pages 1-14, August.
Handle:
RePEc:gam:jmathe:v:12:y:2024:i:17:p:2703-:d:1467458
Download full text from publisher
Corrections
All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:gam:jmathe:v:12:y:2024:i:17:p:2703-:d:1467458. See general information about how to correct material in RePEc.
If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.
We have no bibliographic references for this item. You can help adding them by using this form .
If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: MDPI Indexing Manager (email available below). General contact details of provider: https://www.mdpi.com .
Please note that corrections may take a couple of weeks to filter through
the various RePEc services.