Scholarlink

<p>Eye movement (EM) detection is a critical step in most eye-tracking (ET) research, typically relying on <i>detectors</i>–specialized algorithms designed to segment raw ET data into discrete oculomotor events. However, variability in detection algorithms and the lack of standardized evaluation frameworks hinder transparency and reproducibility across studies. In this work, we introduce <i>pEYES</i>, an open-source toolkit designed to streamline EM detection and enable robust, quantitative comparisons between detectors. The toolkit provides implementations for several widely used threshold-based detectors, along with multiple standardized evaluation procedures for assessing detection performance. Using <i>pEYES</i>, we evaluated seven detection algorithms on two publicly available human-annotated datasets containing recordings of subjects freely viewing color images. Performance was assessed using metrics such as Cohen’s kappa, relative timing offset and deviation, and a sensitivity index (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(d{\prime}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>d</mi> <mo>′</mo> </mrow> </math></EquationSource> </InlineEquation>) for fixation and saccade onsets and offsets. Engbert’s adaptive velocity-threshold algorithm consistently matched or outperformed the other detectors, occasionally achieving human-level precision. In contrast, several other detectors exhibited substantial variability in performance between datasets. We also found systematic differences in detection scores between fixation and saccade boundaries, with fixation offsets and saccade onsets detected more reliably than their counterparts. These findings highlight the importance of task- and dataset-specific detector selection in EM analysis. The <i>pEYES</i> toolkit is freely available, and its codebase&#xa0;–&#xa0;along with the analyses presented in this report&#xa0;–&#xa0;is accessible at <a href="https://github.com/huji-hcnl/pEYES">https://github.com/huji-hcnl/pEYES</a>. We invite the research community to use, extend, and contribute to its ongoing development. Through open collaboration, we aim to advance the rigor and reproducibility of EM detection practices.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Systematic classification differences across eye movement detection algorithms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月10日

查看全文
  • Jonathan Nir,
  • Leon Y. Deouell

摘要

Eye movement (EM) detection is a critical step in most eye-tracking (ET) research, typically relying on detectors–specialized algorithms designed to segment raw ET data into discrete oculomotor events. However, variability in detection algorithms and the lack of standardized evaluation frameworks hinder transparency and reproducibility across studies. In this work, we introduce pEYES, an open-source toolkit designed to streamline EM detection and enable robust, quantitative comparisons between detectors. The toolkit provides implementations for several widely used threshold-based detectors, along with multiple standardized evaluation procedures for assessing detection performance. Using pEYES, we evaluated seven detection algorithms on two publicly available human-annotated datasets containing recordings of subjects freely viewing color images. Performance was assessed using metrics such as Cohen’s kappa, relative timing offset and deviation, and a sensitivity index ( \(d{\prime}\) d ′ ) for fixation and saccade onsets and offsets. Engbert’s adaptive velocity-threshold algorithm consistently matched or outperformed the other detectors, occasionally achieving human-level precision. In contrast, several other detectors exhibited substantial variability in performance between datasets. We also found systematic differences in detection scores between fixation and saccade boundaries, with fixation offsets and saccade onsets detected more reliably than their counterparts. These findings highlight the importance of task- and dataset-specific detector selection in EM analysis. The pEYES toolkit is freely available, and its codebase – along with the analyses presented in this report – is accessible at https://github.com/huji-hcnl/pEYES. We invite the research community to use, extend, and contribute to its ongoing development. Through open collaboration, we aim to advance the rigor and reproducibility of EM detection practices.