Approximations of Pupillary Shape in High-Definition Video-Oculography Registers
Two approximations of the pupillary shape (in laboratory environment) were developed: a nonlinear elliptic fit with outliers rejection and the assessment of the realistic amorphous shape via radial distances measurement, for the more diaphanous registers.
- PDF / 943,152 Bytes
- 4 Pages / 595.276 x 790.866 pts Page_size
- 63 Downloads / 187 Views
Abstract—Two approximations of the pupillary shape (in laboratory environment) were developed: a nonlinear elliptic fit with outliers rejection and the assessment of the realistic amorphous shape via radial distances measurement, for the more diaphanous registers. The present methodology achieved a resolution of 10 μm/pixel. The algorithm of nonlinear elliptic fit with outliers rejection resulted a robust option for infrared reflex artifacts, whereas, radial distance measurement could lead us towards a new way of spatiotemporal analysis of pupillary dynamic. Keywords— Pupil, Video-Oculography, Elliptic Fit, Radial Distances, High-Definition. I. INTRODUCTION
Pupil is a structure located over the iris, which plays the role of a control diaphragm of the amount of light that arrives to the retina. Its contraction (miosis) is effectuated by the sphincter muscle (mediated by the parasympathetic system) and the responsible of its dilatation (mydriasis) is the dilator muscle (sympathetic activity) [1]. The study of pupillary dynamic has been a source of information for a variety of clinical and engineering applications. Some of the pupillary event of interest are the pupil cycle time [2], pupillary latency [3], hippus [4], also in an indirect form, ocular movements [5]. Some examples of clinical applications are studies of Type I-II Diabetes [6], Nervous system diseases [7] etc. whereas engineering application include but are not limited to Biometric systems [4] and Eye-trackers [8]. For the pupillary dynamic assessment, a variety of methodologies (pupilometers) have been developed as well as commercial and for research purposes. The proposed algorithms for this second group solve specific needs, for example, the real-time detection of the pupillary centroid for eye trackers [8] (which means very fast and robust detection), high sample rate for eye movements [9] (which means a high frame rate, in other words, high speed cameras) or high spatial resolution for analysis of differences in consensual and direct pupillary reflex (which means HD cameras) [10]. Typically, pupillary shape is fitted to a circular or elliptic approximation. In [11] more realistic analysis of the pupillary shape was introduced. The present work deals with the most
general case of ellipse fitting (centered out of origin and rotated) besides, an alternative of realistic assessment is performed. The last implementation is based on the measurement of radial distances to the edge of the pupil. II.
METHODOLOGY
For the present experiment, two modules were implemented. The first one involves the development of the control interface of the luminous stimuli and video recording, whereas the second refers to the techniques of digital image processing employed in the measurement of the pupillary characteristics. The next sections describe, in detail, all particularities of the two modules. All registers were processed off-line in the Matlab platform. As it was early mentioned, all registers were taken in a laboratory environment. The room for the tests was cond
Data Loading...
