SSBC 2015: Sclera segmentation benchmarking competition

Abstract

This article reports and summarizes the results of a competition on sclera segmentation benchmarking, called Sclera Segmentation Benchmarking Competition 2015 (SSBC 2015). It was organized in the context of the Seventh IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS 2015). The goal of this competition was to set a common platform for evaluation of the recent development in sclera segmentation and also to attract the attention of researchers to this new area. In this regard, we have prepared a Multi- Angle Sclera Dataset (MASD) version 1. It is comprised of 2624 RGB images taken from both the eyes of 82 identities. In other words, it consists of images of 164 (82*2) different eyes. Moreover, we have prepared manual segmentation masks of these images to create the baseline. We have, furthermore, adopted precision and recall- based statistical measures to evaluate the effectiveness and the ranks of the competing algorithms. To summarize, ten participants registered for the competition, and among them four participants submitted their algorithms/systems. The results achieved by these algorithms reflect noteworthy developments in the literature of sclera segmentation employing cutting edge segmentation techniques. Along with the algorithm of the four competing teams and their results, the MASD version 1 dataset will also be freely available for research purposes from the organizer's website.