Authors:
Jiali Cui
;
Fuqiang Chen
;
Duo Shi
and
Liqiang Liu
Affiliation:
North China University of Technology, China
Keyword(s):
Iris recognition, Eye detection, Deep convolutional neural networks, Faster R-CNN.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Computational Intelligence
;
Evolutionary Computing
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Machine Learning
;
Soft Computing
;
Symbolic Systems
Abstract:
The accuracy of eye detection is crucial to a variety of biometric identification technologies, such as iris recognition. The challenge of eye detection comes from improving accuracy of detection in the case of occlusion or reflection of glasses. In this paper, an eye detection method based on Faster Region-based Convolutional Neural Network (Faster R-CNN) is proposed. The method includes three import parts: convolutional layers, region proposal network (RPN) and detection network. By training monocular and binocular models on the training dataset, the recall of monocular and binocular models on test dataset can reach 96% and 95% respectively, which proves that the proposed method based on Faster R-CNN has high accuracy of detection in the task.