arXiv subjects

Jaka Sodnik

Publications and source records attributed to Jaka Sodnik.

5 recordsLinked to original sources

arXiv2025

From Raw Gaze to Meaningful Features: Assessment of Visual Behavior in Driving Simulator

A novel quantitative analytical approach to investigate eye movement behaviour in a driving simulator during three conditions: Baseline, Ride (simulated drive under normal visibility), and Fog (simulated drive under reduced visibility) is presented. We illustrate the proposed approach in a case study: eye tracking data from 24 participants are analyzed using 31 parameters, organized into three groups: (1) saccade characteristics, (2) Bivariate Contour Ellipse Area (BCEA) related parameters, and (3) blinking features. Across all feature groups, numerous statistically significant differences emerge between Baseline and the simulated drive conditions. Between Ride and Fog, saccade features show minimal changes (one out of 13), whereas BCEA (9 of 13) and blink features (four of 5) exhibit pronounced differences, highlighting the strong impact of reduced visibility on gaze stability and blinking behaviour. In addition to conventional gaze features, such as saccade duration, a new metric, inspired by the Guzik's Index (GI), is introduced through feature engineering to quantify fixation asymmetry along the major axis of the BCEA. The GI results suggest that gaze dispersion remains consistent across visibility conditions, but it is significantly higher during Baseline than during Fog and Ride conditions (p < 0.01). Overall, the presented findings demonstrate that a combination of the BCEA-based features with saccade and blink parameters provides a comprehensive understanding of visual attention and gaze stability in a driving simulator, while GI offers additional insights into fixation asymmetry under varying visibility conditions.

eess.SP↗

arXiv2025

Contactless pulse rate assessment: Results and insights for application in driving simulator

Remote photoplethysmography (rPPG) offers a promising solution for non-contact driver monitoring by detecting subtle blood flow-induced facial color changes from video. However, motion artifacts in dynamic driving environments remain key challenges. This study presents an rPPG framework that combines signal processing techniques before and after applying Eulerian Video Magnification (EVM) for pulse rate (PR) estimation in driving simulators. While not novel, the approach offers insights into the efficiency of the EVM method and its time complexity. We compare results of the proposed rPPG approach against reference Empatica E4 data and also compare it with existing achievements from the literature. Additionally, the possible bias of the Empatica E4 is further assessed using an independent dataset with both the Empatica E4 and the Faros 360 measurements. EVM slightly improves PR estimation, reducing the mean absolute error (MAE) from 6.48 bpm to 5.04 bpm (the lowest MAE (~2 bpm) was achieved under strict conditions) with an additional time required for EVM of about 20 s for 30 s sequence. Furthermore, statistically significant differences are identified between younger and older drivers in both reference and rPPG data. Our findings demonstrate the feasibility of using rPPG-based PR monitoring, encouraging further research in driving simulations.

eess.IV↗

arXiv2024

Data Augmentation for Generating Synthetic Electrogastrogram Time Series

To address an emerging need for large number of diverse datasets for rigor evaluation of signal processing techniques, we developed and evaluated a new method for generating synthetic electrogastrogram time series. We used electrogastrography (EGG) data from an open database to set model parameters and statistical tests to evaluate synthesized data. Additionally, we illustrated method customization for generating artificial EGG time series alterations caused by the simulator sickness. Proposed data augmentation method generates synthetic EGG data with specified duration, sampling frequency, recording state (postprandial or fasting state), overall noise and breathing artifact injection, and pauses in the gastric rhythm (arrhythmia occurrence) with statistically significant difference between postprandial and fasting states in > 70% cases while not accounting for individual differences. Features obtained from the synthetic EGG signal resembling simulator sickness occurrence displayed expected trends. The code for generation of synthetic EGG time series is not only freely available and can be further customized to assess signal processing algorithms but also may be used to increase data diversity for training artificial intelligence (AI) algorithms. The proposed approach is customized for EGG data synthesis but can be easily utilized for other biosignals with similar nature such as electroencephalogram.

eess.SP↗

arXiv2023

Effectiveness of a time to fixate for fitness to drive evaluation in neurological patients

We present a method to automatically calculate time to fixate (TTF) from the eye-tracker data in subjects with neurological impairment using a driving simulator. TTF presents the time interval for a person to notice the stimulus from its first occurrence. Precisely, we measured the time since the children started to cross the street until the drivers directed their look to the children. From 108 neurological patients recruited for the study, the analysis of TTF was performed in 56 patients to assess fit-, unfit-, and conditionally-fit-to-drive patients. The results showed that the proposed method based on the YOLO (you only look once) object detector is efficient for computing TTFs from the eye-tracker data. We obtained discriminative results for fit-to-drive patients by application of Tukey's honest significant difference post hoc test (p < 0.01), while no difference was observed between conditionally-fit and unfit-to-drive groups (p = 0.542). Moreover, we show that time-to-collision (TTC), initial gaze distance (IGD) from pedestrians, and speed at the hazard onset did not influence the result, while the only significant interaction is among fitness, IGD, and TTC on TTF. Obtained TTFs are also compared with the perception response times (PRT) calculated independently from eye-tracker data and YOLO. Although we reached statistically significant results that speak in favor of possible method application for assessment of fitness to drive, we provide detailed directions for future driving simulation-based evaluation and propose processing workflow to secure reliable TTF calculation and its possible application in for example psychology and neuroscience.

eess.IV↗

arXiv2021

Driver2vec: Driver Identification from Automotive Data

With increasing focus on privacy protection, alternative methods to identify vehicle operator without the use of biometric identifiers have gained traction for automotive data analysis. The wide variety of sensors installed on modern vehicles enable autonomous driving, reduce accidents and improve vehicle handling. On the other hand, the data these sensors collect reflect drivers' habit. Drivers' use of turn indicators, following distance, rate of acceleration, etc. can be transformed to an embedding that is representative of their behavior and identity. In this paper, we develop a deep learning architecture (Driver2vec) to map a short interval of driving data into an embedding space that represents the driver's behavior to assist in driver identification. We develop a custom model that leverages performance gains of temporal convolutional networks, embedding separation power of triplet loss and classification accuracy of gradient boosting decision trees. Trained on a dataset of 51 drivers provided by Nervtech, Driver2vec is able to accurately identify the driver from a short 10-second interval of sensor data, achieving an average pairwise driver identification accuracy of 83.1% from this 10-second interval, which is remarkably higher than performance obtained in previous studies. We then analyzed performance of Driver2vec to show that its performance is consistent across scenarios and that modeling choices are sound.

cs.LG↗