arXiv subjects

Benedikt W. Hosp

Publications and source records attributed to Benedikt W. Hosp.

4 recordsLinked to original sources

arXiv2025

FOVAL: Calibration-Free and Subject-Invariant Fixation Depth Estimation Across Diverse Eye-Tracking Datasets

Accurate fixation depth estimation is essential for applications in extended reality (XR), robotics, and human-computer interaction. However, current methods heavily depend on user-specific calibration, which limits their scalability and usability. We introduce FOVAL, a robust calibration-free approach that combines spatiotemporal sequence modelling via Long Short-Term Memory (LSTM) networks with subject-invariant feature engineering and normalisation. Compared to Transformers, Temporal Convolutional Networks (TCNs), and CNNs, FOVAL achieves superior performance, particularly in scenarios with limited and noisy gaze data. Evaluations across three benchmark datasets using Leave-One-Out Cross-Validation (LOOCV) and cross-dataset validation show a mean absolute error (MAE) of 9.1 cm and strong generalisation without calibration. We further analyse inter-subject variability and domain shifts, providing insight into model robustness and adaptation. FOVAL's scalability and accuracy make it highly suitable for real-world deployment.

cs.CV↗

arXiv2025

DeepACTIF: Efficient Feature Attribution via Activation Traces in Neural Sequence Models

Feature attribution is essential for interpreting deep learning models, particularly in time-series domains such as healthcare, biometrics, and human-AI interaction. However, standard attribution methods, such as Integrated Gradients or SHAP, are computationally intensive and not well-suited for real-time applications. We present DeepACTIF, a lightweight and architecture-aware feature attribution method that leverages internal activations of sequence models to estimate feature importance efficiently. Focusing on LSTM-based networks, we introduce an inverse-weighted aggregation scheme that emphasises stability and magnitude of activations across time steps. Our evaluation across three biometric gaze datasets shows that DeepACTIF not only preserves predictive performance under severe feature reduction (top 10% of features) but also significantly outperforms established methods, including SHAP, IG, and DeepLIFT, in terms of both accuracy and statistical robustness. Using Wilcoxon signed-rank tests and effect size analysis, we demonstrate that DeepACTIF yields more informative feature rankings with significantly lower error across all top-k conditions (10 - 40%). Our experiments demonstrate that DeepACTIF not only reduces computation time and memory usage by orders of magnitude but also preserves model accuracy when using only top-ranked features. That makes DeepACTIF a viable solution for real-time interpretability on edge devices such as mobile XR headsets or embedded health monitors.

cs.LG↗

arXiv2023

How to Tune Autofocals: A Comparative Study of Advanced Tuning Methods

This study comprehensively evaluates tuning methods for autofocal glasses using virtual reality (VR), addressing the challenge of presbyopia. With aging, presbyopia diminishes the eye's ability to focus on nearby objects, impacting the quality of life for billions. Autofocals, employing focus-tunable lenses, dynamically adjust optical power for each fixation, promising a more natural visual experience than traditional bifocal or multifocal lenses. Our research contrasts the most common tuning methods - manual, gaze-based, and vergence - within a VR setup to mimic real-world scenarios. Utilizing the XTAL VR headset equipped with eye-tracking, the study replicated autofocal scenarios, measuring performance and usability through psychophysical tasks and NASA TLX surveys. Results show varying strengths and weaknesses across methods, with gaze control excelling in precision but not necessarily comfort and manual control providing stability and predictability. The findings guide the selection of tuning methods based on task requirements and user preferences, highlighting a balance between precision and ease of use.

cs.HC↗

arXiv2023

VisionaryVR: An Optical Simulation Tool for Evaluating and Optimizing Vision Correction Solutions in Virtual Reality

Developing and evaluating vision science methods require robust and efficient tools for assessing their performance in various real-world scenarios. This study presents a novel virtual reality (VR) simulation tool that simulates real-world optical methods while giving high experimental control to the experiment. The tool incorporates an experiment controller, to smoothly and easily handle multiple conditions, a generic eye-tracking controller, that works with most common VR eye-trackers, a configurable defocus simulator, and a generic VR questionnaire loader to assess participants' behavior in virtual reality. This VR-based simulation tool bridges the gap between theoretical and applied research on new optical methods, corrections, and therapies. It enables vision scientists to increase their research tools with a robust, realistic, and fast research environment.

cs.CV↗