Gaze prediction is essential for addressing motion-to-photon latency and ensuring seamless foveated rendering in Virtual Reality. The reliability of gaze forecasting is highly sensitive to individual differences and the eye movements being predicted. We evaluate recurrent, transformer-based, and classification-guided architectures to assess their generalization capabilities across oculomotor events. Using the GazeBase VR and Meta Quest Pro datasets, we analyzed the relationship between the median (P50) and high-percentile (P95) error profiles across subjects. The analysis reveals significant performance variability, showing that subjects with low P50 errors do not always exhibit the lowest extreme-case errors. Consequently, low median errors do not guarantee the robustness of the utilized solution. We discuss inference performance and address the class imbalance problem in short-term gaze prediction. These results identify a gap in standardized evaluation methods, necessitating a shift toward P95-focused, subject-specific metrics to develop reliable and perceptually stable gaze-contingent systems.

Eye movement biometric has recently reached a meaningful performance threshold within a gaze estimation pipeline. Prior research claimed that good biometric performance can be achieved from a relatively large set of weakly intercorrelated features with high temporal persistence (indexed by the measurement of reliability). In this study, we revisit this hypothesis in the context of a modern deep learning (DL)-based eye movement biometric system, using a publicly available eye-movement dataset. Specifically, we investigate whether the measurement of reliability of learned embeddings continues to predict biometric performance in DL-based biometrics. Our results confirm that temporal persistence—quantified by measurement of reliability—is a significant predictor of performance in DL-based biometric systems, extending prior findings into the DL-based biometric. We also examine how manipulating eye-tracking signal quality descriptors impacts the temporal persistence of embeddings, finding that degradation of any kind undermines their temporal persistence. As a general matter, we found that measurement of reliability is an important predictor of DL-based biometric performance, and also that DL-learned embeddings are generally weakly intercorrelated.

We present an analysis of the eye tracking capabilities of the Meta Quest Pro virtual reality headset using a dataset of eye movement recordings collected from 78 participants. We highlight the potential differences in user experience as a function of device performance using a novel, user-centered evaluation framework for eye tracking signal quality analysis. In addition to presenting classical signal quality metrics such as spatial accuracy and spatial precision, we also explore how spatial accuracy varies across the field of view for different users across the performance range of the device. This work contributes to a growing understanding of eye tracking signal quality in virtual reality platforms, where the usability of gaze-based applications is directly dependent on the quality of the device’s eye tracking signal.

Prior research states that frequencies below 75 Hz in eye-tracking data represent the primary eye movement termed “signal” while those above 75 Hz are deemed “noise”. This study examines the biometric significance of this signal-noise distinction and its privacy implications. There are important individual differences in a person’s eye movement, which lead to reliable biometric performance in the “signal” part. Despite minimal eye-movement information in the “noise” recordings, there might be significant individual differences. Our results confirm the “signal” predominantly contains identity-specific information, yet the “noise” also possesses unexpected identity-specific data. This consistency holds for both short-(≈ 20 min) and long-term (≈ 1 year) biometric evaluations. Understanding the location of identity data within the eye movement spectrum is essential for privacy preservation.

The importance of normalizing biometric features or matching scores is understood in the multimodal biometric case, but there is less attention to the unimodal case. Prior reports assess the effectiveness of normalization directly on biometric performance. We propose that this process is logically comprised of two independent steps: (1) methods to equalize the effect of each biometric feature on the similarity scores calculated from all the features together and (2) methods of weighting the normalized features to optimize biometric performance. In this report, we address step 1 only and focus exclusively on normally distributed features. We show how differences in the variance of features lead to differences in the strength of the influence of each feature on the similarity scores produced from all the features. Since these differences in variance have nothing to do with importance in the biometric sense, it makes no sense to allow them to have greater weight in the assessment of biometric performance. We employed two types of features: (1) real eye-movement features and (2) synthetic features. We compare six variance normalization methods (histogram equalization, L1-normalization, median normalization, z-score normalization, min-max normalization, and L-infinite normalization) and one distance metric (Mahalanobis distance) in terms of how well they reduce the impact of the variance differences. The effectiveness of different techniques on real data depended on the strength of the inter-correlation of the features. For weakly correlated real features and synthetic features, histogram equalization was the best method followed by L1 normalization.