An accurate and fast method for eyelid detection
by Ahmed AK Tahir; Steluta Anghelus
International Journal of Biometrics (IJBM), Vol. 12, No. 2, 2020

Abstract: A novel method called refine-connect-extend-smooth (R-C-E-S) for detecting eyelids is presented. It consists of four algorithms, Canny edge detector with Prewitt operator, modified refine edge map (MREM), connect edges-extend (CEE) and smooth curve (SC). The method is not based on pre-assumptions that consider eyelids as parabola or lines and it does not use curve fitting algorithm, therefore sever deviation of the detected eyelid curve from the actual eyelid path is avoided. The method is applied to three types of database, CASIA-V1.0, CASIA-V4.0-Lamp and SDUMLA-HMT. The accuracies for detecting the lower eyelid, upper eyelid and free iris are (93.2%, 99.1%, 96.7%) for CASIA-V1.0, while for CASIA-V4.0-Lamp are (97.6%, 98.3%, 97.8%) and for SDUMLA-HMT are (95.1%, 95.3%, 96.92%). The processing times for detecting single eyelid, both eyelids and free iris are (42 ms, 49 ms, 35 ms) for CASIA-V1.0, while for CASIA-V4.0-Lamp are (23 ms, 26 ms, 21 ms) and for SDUMLA-HMT are (35 ms, 40 ms, 31 ms).

Online publication date: Wed, 10-Jun-2020

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Biometrics (IJBM):
Login with your Inderscience username and password:

    Username:        Password:         

Forgotten your password?


Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com