What does this research mean for the field?

An image-based framework that represents eye movement trajectories as grayscale images provides a complementary analysis of gaze behavior that captures both conventional fixation metrics and spatial characteristics without requiring predefined Area-of-Interest boundaries. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To develop and validate an image-based statistical framework that analyzes the spatial organization of eye movement trajectories without requiring predefined area-of-interest boundaries.

September 15, 2026MathematicsOpen Access

Statistical Image Analysis of Eye Movement Trajectories in Multi-Attribute Decision Making

View Full Paper

Ask AI

Bookmark

Authors

KTKazuhisa Takemura KKKeita Kawasugi HMH. Murakami

Discussion

Member takes

Overview

Methodological study demonstrates image-based eye-tracking trajectory modeling in multi-attribute decision tasks, highlighting spatial patterns beyond traditional region-of-interest metrics.

Key Points

To develop and validate an image-based statistical framework that analyzes the spatial organization of eye movement trajectories without requiring predefined area-of-interest boundaries.
Converted eye-tracking trajectories from participants performing multi-attribute insurance decision tasks into grayscale images on a regular spatial grid.
Extracted spatial features using texture analysis, Fourier analysis, wavelet decomposition, and singular value decomposition, excluding non-negative matrix factorization due to limited contribution.
Integrated heterogeneous image features using a Tucker-1 framework and evaluated latent dimensions via principal component analysis and probabilistic principal component analysis with bootstrap validation.
The first principal component demonstrated strong, consistent correlations with traditional fixation count and fixation duration across both insurance tasks, reflecting global fixation intensity.
Higher-order principal components showed substantially weaker correlations with traditional fixation metrics, capturing unique spatial and structural trajectory characteristics.
Bootstrap analyses supported a relatively high-dimensional latent structure, though the exact dimensionality derived from probabilistic principal component analysis was sensitive to sample variability.

Cite This Study

Takemura et al. (2026) studied this question.

synapsesocial.com/papers/6aa913819013453be30a16e7 https://doi.org/10.3390/math14183316

View Full Paper

Ask AI

Bookmark

September 15, 2026MathematicsOpen Access

Statistical Image Analysis of Eye Movement Trajectories in Multi-Attribute Decision Making

View Full Paper

Ask AI

Bookmark

Authors

KTKazuhisa Takemura KKKeita Kawasugi HMH. Murakami

Discussion

Member takes

Overview

Methodological study demonstrates image-based eye-tracking trajectory modeling in multi-attribute decision tasks, highlighting spatial patterns beyond traditional region-of-interest metrics.

Key Points

To develop and validate an image-based statistical framework that analyzes the spatial organization of eye movement trajectories without requiring predefined area-of-interest boundaries.
Converted eye-tracking trajectories from participants performing multi-attribute insurance decision tasks into grayscale images on a regular spatial grid.
Extracted spatial features using texture analysis, Fourier analysis, wavelet decomposition, and singular value decomposition, excluding non-negative matrix factorization due to limited contribution.
Integrated heterogeneous image features using a Tucker-1 framework and evaluated latent dimensions via principal component analysis and probabilistic principal component analysis with bootstrap validation.
The first principal component demonstrated strong, consistent correlations with traditional fixation count and fixation duration across both insurance tasks, reflecting global fixation intensity.
Higher-order principal components showed substantially weaker correlations with traditional fixation metrics, capturing unique spatial and structural trajectory characteristics.
Bootstrap analyses supported a relatively high-dimensional latent structure, though the exact dimensionality derived from probabilistic principal component analysis was sensitive to sample variability.

Cite This Study

Takemura et al. (2026) studied this question.

synapsesocial.com/papers/6aa913819013453be30a16e7 https://doi.org/10.3390/math14183316

View Full Paper

Ask AI

Bookmark