Raimondas Zemblys

Correction to: “Using machine learning to detect events in eye-tracking data”

It has come to our attention that the section “Post-processing: Labeling final events” on page 167 of “Using Machine Learning to Detect Events in Eye-Tracking Data” (Zemblys, Niehorster, Komogortsev, & Holmqvist, 2018) contains an erroneous description of the process by which post-processing was performed.

Evaluating Informational Characteristics of Oculomotor System and Human–Computer Interfaces

Interest in adapting eye movements to everyday interaction is apparent in recent consumer technology: more and more interfaces with gaze as input will be developed in years to come. Usually throughput and other Fitts’ index of difficulty based methods are used to evaluate human’s ability to perform a coordinated movement. However, it is questionable if such methods apply to gaze based interaction. This paper explores measures of information transfer rate and channel capacity—a better alternative to throughput. Based on difference of initial and final entropy of the extent of the movement, these measures can be used to evaluate and compare any pointing devices (mouse, joystick, eye tracker, etc.) because they evaluate only information processing capacity of oculomotor channel, without including informational characteristics of the methods to perform object selection.

EFFECTS OF MÜLLER-LYER ILLUSION ON THE ACCURACY OF PRIMARY SACCADES AND SMOOTH PURSUIT EYE MOVEMENTS

The goal of the present investigation was to evaluate the effect of visual Muller-Lyer (M-L) illusion on the control of saccadic and smooth pursuit eye movements and compare it with the results obtained during perceptual judgment of the length of the shaft of M-L illusion. Experimental investigation revealed that the primary saccades elicited in the reflexive mode were mostly affected by the M-L illusion. The position errors of the primary saccades elicited in the reflexive mode were affected by 4% for wings-in illusion and by 3.6% for wings-out illusion comparing with the 0.25% and 0.1% for the saccades elicited in the voluntary mode. The position errors of complete saccades (0.14% and 0.02%) and tracking errors obtained during the smooth pursuit (0.11% and 0.05%) were negligibly small. Nevertheless, experimental results obtained during perceptual judgment of M-L illusion were substantially larger - 14% and 10% respectively. The obtained results have demonstrated that vision for perception and vision for action were executed by different neural networks and supported the two-visual-system hypothesis. DOI: http://dx.doi.org/10.5755/j01.itc.41.1.763

Influence of successive catch-up saccades on the accuracy of tracking eye movements

Catch-up saccades and smooth pursuit are two components, quick and slow, of tracking eye movements. Their interaction has usually been studied during tracking a target moving only in the horizontal direction. In these experiments catch-up saccades were invoked by changing target velocity or introducing target jumps and occlusions. Our research was focused on ocular pursuit of two-dimensional non-predictive target trajectories without target jumps. Target trajectories presented to the subjects had the same shape but moved with slow, medium and high average speeds, and each trial had the duration of 60, 30 and 15 s, respectively. During the long tracking experiments, the sequences of successive catch-up saccades were obtained and their parameters were computed. Due to the contribution of catch-up saccades, elicited during ocular pursuit, the position errors and lagging times were reduced from 30 to 50 % and larger values were obtained for higher target velocities. Analyzing the sum of tracking distances which were overcome during slow phases and during catch-up saccades, we proposed new method to evaluate subjects pursuit gain.

Modeling of the manuo-ocular coordination during object guiding through a path

Knowledge on human eye-hand coordination can be used for human-like system design and medical diagnosis. This document analyses and briefly presents the parameters of the coordination while executing different eye-hand related tasks. Existing quantitative model of manuo-ocular coordination, capable of simulating the human performance in target tracking, is redesigned for a capability to simulate the performance of object guiding through a visible predefined path. Qualitative model, based on quantitative model, is proposed and explained.

Quantitative Analysis of Two-Dimensional Catch-Up Saccades Executed to the Target Jumps in the Time-Continuous Trajectory

The purpose of this research was to investigate quantitatively the catch-up saccades occurring during smooth pursuit. In the first experiment, to evoke catch-up saccades we used high velocity predictable two-dimensional time-continuous target trajectories. In the second experiment, catch-up saccades were evoked using target jump paradigm during sustained two-dimensional pursuit. Target jumps in the different directions were presented at the unexpected moments and positions of the interrupted time-continuous target trajectory. From the experimental results we made a comparison of the main sequences (relationship between peak velocity and amplitude) of the catch-up and refixation saccades and found that they are different. Also we can conclude that the peak velocity of catch-up saccades is strongly correlated with the velocity of the smooth pursuit target component. We found that both position error and retinal slip are taken into account in catch-up saccades programming to predict the future trajectory of the moving target. Ill. 5, bibl. 5 (in English, summaries in English, Russian and Lithuanian).