Translator Disclaimer
22 September 2015 Correlation peak analysis applied to a sequence of images using two different filters for eye tracking model
Author Affiliations +
Eye tracking has many useful applications that range from biometrics to face recognition and human-computer interaction. The analysis of the characteristics of the eyes has become one of the methods to accomplish the location of the eyes and the tracking of the point of gaze. Characteristics such as the contrast between the iris and the sclera, the shape, and distribution of colors and dark/light zones in the area are the starting point for these analyses. In this work, the focus will be on the contrast between the iris and the sclera, performing a correlation in the frequency domain. The images are acquired with an ordinary camera, which with were taken images of thirty-one volunteers. The reference image is an image of the subjects looking to a point in front of them at 0° angle. Then sequences of images are taken with the subject looking at different angles. These images are processed in MATLAB, obtaining the maximum correlation peak for each image, using two different filters. Each filter were analyzed and then one was selected, which is the filter that gives the best performance in terms of the utility of the data, which is displayed in graphs that shows the decay of the correlation peak as the eye moves progressively at different angle. This data will be used to obtain a mathematical model or function that establishes a relationship between the angle of vision (AOV) and the maximum correlation peak (MCP). This model will be tested using different input images from other subject not contained in the initial database, being able to predict angle of vision using the maximum correlation peak data.
© (2015) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Verónica A. Patrón, Josué Álvarez Borrego, and Ángel Coronel Beltrán "Correlation peak analysis applied to a sequence of images using two different filters for eye tracking model", Proc. SPIE 9599, Applications of Digital Image Processing XXXVIII, 95992R (22 September 2015);


Eyebrow segmentation using active shape models
Proceedings of SPIE (May 31 2013)
Refinement of a model for predicting perceived brightness
Proceedings of SPIE (December 28 2001)
Recognizing human eyes
Proceedings of SPIE (September 01 1991)

Back to Top