Real-time gaze tracking with appearance-based models

Abstract

Psychological evidence has emphasized the importance of eye gaze analysis in human computer interaction and emotion interpretation. To this end, current image analysis algorithms take into consideration eye-lid and iris motion detection using colour information and edge detectors. However, eye movement is fast and and hence difficult to use to obtain a precise and robust tracking. Instead, our method proposed to describe eyelid and iris movements as continuous variables using appearance-based tracking. This approach combines the strengths of adaptive appearance models, optimization methods and backtracking techniques. Thus, in the proposed method textures are learned on-line from near frontal images and illumination changes, occlusions and fast movements are managed. The method achieves real-time performance by combining two appearance-based trackers to a backtracking algorithm for eyelid estimation and another for iris estimation. These contributions represent a significant advance towards a reliable gaze motion description for HCI and expression analysis, where the strength of complementary methodologies are combined to avoid using high quality images, colour information, texture training, camera settings and other time-consuming processes.

Statistically Learned Deformable Eye Models

“…CLM appear to be consistently inferior to AAM and SDM. It is worth noting, that all methods are capable of fitting the sclera, iris and pupil parts accurately and the reported errors are driven by the top (landmarks: [1][2][3][4][5]) and bottom landmarks (landmarks: [7][8][9][10][11]) of the upper and lower eyelids.…”

“…The two closest works to the approach present in this paper are the ones of Moriyama et al [4] and Orozco et al [9]. The authors of [4] propose a 2D handlycrafted parametrised generative eye shape model inspired by the anatomical structure of the human eye.…”

“…On the other hand, the authors of [9] propose an on-line appearance-based tracker that automatically adapts to changes in eye texture. They use a parametrised shape model based on a hand-crafted standard designed by the computer animation industry (Face Animation Parameters (FAP)).…”

“…However, despite recent advances [6,7,8,4,9], accurate and robust eye alignment in unconstrained scenarios remains an extremely challenging task. The main difficulty arises from the very diverse appearance of eyes caused by both anatomical differences between individuals (Figure 1a) and the high deformability and fast movement of the different eye components (Figure 1b).…”

“…A largely diverse number of approaches, ranging from simple techniques based on the application of edge detectors and Hough transform [6,7] to more sophisticated model-based approaches [9,4], have been used to solve the eye alignment problem.…”