A literature survey on eye corner detection techniques in real-life scenarios

Abstract

Accurate iris movement detection and tracking is an important and widely used step in many Human-computer interactive applications. Among the eye features, eye corners are considered as stable and reliable reference points to measure the relative iris motion. In real time scenarios, the presence of spectacles prohibit the current state-of-the-art methods to yield accurate detection as the appearance of eye corners changes considerably due to the glare and occlusion caused by them. We term this problem as the Spectacle problem. In this paper we review the available single and multiple image based spectacle problem removal techniques and highlight the pros and cons of the approaches. For this state-of-the-art report, we investigated research papers, patents and thesis presenting the basic definitions, terminologies and new directions for future researches.