Recognition of understanding level and language skill using measurements of reading behavior

Abstract

The reading act is an intimate and elusive process that is important to understand. Psycholinguists have long studied the effects of task, personal or document characteristics on reading behavior. An essential factor in the success of those studies lies in the capability of analyzing eye-movements. These studies aim to recognize causal effects on patterns of eye-movements, by contriving variations in task, personal or document characteristics. In this work, we follow the opposite direction. We present a formal framework to recognize reader's level of understanding and language skill given measurements of reading behavior via eye-gaze data. We show significant error reductions to recognize these attributes and provide a detailed study of the most discriminative features.

Vertical Error Correction Using Classification of Transitions between Sequential Reading Segments

“…[9] because it features natural reading activity. Eleven participants read three English documents each, for the total of 33 gaze datasets.…”

“…According to Ref. [9], participants read the text naturally, and did not use any equipment to fix their head position.…”

“…According to Martínez-Gómez et al [9], natural reading that normally happens in uncontrolled environments is very hard to process with current equipment, particularly due to the vertical error. Increasing line spacing to avoid the vertical error dramati-cally increases the accuracy of the measurement, but it forces the subjects to read unnaturally.…”

“…Although the results showed improvements in fixation re-mapping, this method does not overcome accumulated systematic errors. Martínez-Gómez et al [9] proposed a feature-based image-registration method that spatially transforms the image representation of fixations to match the image representation of the text. This method uses several optimization strategies to search for the optimal scaling and translation parameters.…”

Spatio-Temporal Detection of Divided Attention in Reading Applications Using EEG and Eye Tracking

“…Classification was reported to be on average 76% accurate (55%-60% for the inattention class, 87%-90% for attention, depending on the cost function used). Martínez-Gómez et al [11] found characteristic features from eye tracking data that describe a subject's level of understanding and English language skill. Putze et al [13] used the combination of EEG and eye tracking to facilitate video surveillance event selection.…”