Ways of improving the precision of eye tracking data: Controlling the influence of dirt and dust on pupil detection

Abstract

Eye-tracking technology has to date been primarily employed in research. With recent advances in aordable video-based devices, the implementation of gaze-aware smartphones, and marketable driver monitoring systems, a considerable step towards pervasive eye-tracking has been made. However, several new challenges arise with the usage of eye-tracking in the wild and will need to be tackled to increase the acceptance of this technology. The main challenge is still related to the usage of eye-tracking together with eyeglasses, which in combination with reflections for changing illumination conditions will make a subject "untrackable". If we really want to bring the technology to the consumer, we cannot simply exclude 30% of the population as potential users only because they wear eyeglasses, nor can we make them clean their glasses and the device regularly. Instead, the pupil detection algorithms need to be made robust to potential sources of noise. We hypothesize that the amount of dust and dirt on the eyeglasses and the eye-tracker camera has a significant influence on the performance of currently available pupil detection algorithms. Therefore, in this work, we present a systematic study of the eect of dust and dirt on the pupil detection by simulating various quantities of dirt and dust on eyeglasses. Our results show 1) an overall high robustness to dust in an o-focus layer. 2) the vulnerability of edge-based methods to even small in-focus dust particles. 3) a trade-o between tolerated particle size and particle amount, where a small number of rather large particles showed only a minor performance impact.

HPCGen: Hierarchical K-Means Clustering and Level Based Principal Components for Scan Path Genaration

“…In self-diagnostic systems, a disease or defect is diagnosed based on the user's behavior [62,81,65,10,74,17], which can then later be treated or examined by a doctor. For these systems to work reliably, it is also necessary to use a lot of data and many different features like the eye lid closere [47,46,48], pupil variations [41,52,56,51,50,28] or other features extracted from the eye [54,39,27,25,23,55,26,44,22]. In the area of driver observation, the same applies.…”

GroupGazer: A Tool to Compute the Gaze per Participant in Groups with integrated Calibration to Map the Gaze Online to a Screen or Beamer Projection

“…This involves not only simple control but also collaboration in which a human communicates complex behavior to a robot or system [95]. Interaction with the eyes based on pupil movements [56,41,29,27,25,57] is also an interesting source of information in the field of computer games [2]. Eye interaction is many times faster than mouse interaction, which could revolutionize the professional computer gaming field [73].…”

Tensor Normalization and Full Distribution Training

“…Deep neural networks are the state of the art in many areas of image processing. The application fields are image classification [23,25,47,48,33,8,9,18,20,21], semantic segmentation [15,27,11], landmark regression [45,19,49], object detection [51,17,34,24,16,14,12,36], and many more. In the real world, this concerns autonomous driving, human-machine interaction [38,10,35], eye tracking [7,30,29,31,26,32,53,52,13,50], robot control, facial recognition, medical diagnostic systems, and many other areas [39,28,46].…”