Accurate CNN-based pupil segmentation with an ellipse fit error regularization term. (February 2022)

• Record Type: Journal Article
• Authors: Akinlar, Cuneyt; Kucukkartal, Hatice Kubra; Topal, Cihan
• Is Part Of: Expert systems with applications. Volume 188(2022)
• Publication Date: 2022-02
• Subjects: Pupil segmentation -- Convolutional Neural Networks (CNN) -- UNet -- Loss function -- Regularization term; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2021.116004
• Languages: English
• ISSNs: 0957-4174

Abstract

Abstract: Semantic segmentation of images by Fully Convolutional Neural Networks (FCN) has gained increased attention in recent years as FCNs greatly outperform traditional segmentation algorithms. In this paper we propose using Ellipse Fit Error as a shape prior regularization term that can be added to a pixel-wise loss function, e.g., binary cross entropy, to train a CNN for pupil segmentation. We evaluate the performance of the proposed method by training a lightweight UNet architecture, and use three widely used real-world datasets for pupil center estimation, i.e., ExCuSe, ElSe, and Labeled Pupils in the Wild (LPW), containing a total of ∼ 230.000 images for performance evaluation. Experimental results show that the proposed method gives the best-known pupil detection rates for all datasets. Highlights: Proposes Ellipse Fit Error as a regularization term for CNN-based pupil segmentation. Reports the best-known pupil detection results for ExCuSe & LPW datasets. Gives out GT Segmentation maps for the LPW dataset.