An immersive (V)irtual/(A)ugmented Reality experience requires high-resolution display capability and contextual knowledge of user’s intention to customize the content and display elements. Both these tasks critically depend on highly accurate eye-tracking solutions, ideally, ahead of time. Unfortunately, the current generation of eye-tracking algorithms neither meet the exacting accuracy/precision demands nor can they reliably predict the eye-movements. Therefore, we need to push the boundaries of the current just-in-time eye-tracking solutions in terms of a) accuracy and b) eye-movement prediction capabilities.
The goal for this workshop is to motivate the broader community of computer vision and machine learning scientists to actively participate in developing novel outlooks towards eye-tracking solutions to help reach the aforementioned goals. Partly, this workshop is also motivated from the tremendous success of OpenEDS 2019, hosted at ICCV’19, which witnessed enthusiastic participation from researchers in different disciplines such as computer vision, machine learning, eye-tracking and computer graphics. Continuing with the theme of academic engagement, this year, the workshop will focus on the emergence of novel sensor technology for eye-tracking and the wide-scale adoption of eye-tracking in applications ranging from AR, VR, gaming and health. This workshop will host two challenges structured around fusing 2D-3D information and eye-movement prediction research.
These challenges are also accompanied with a release of relevant datasets to support the advancement of research. More information about these challenges can be found on the OpenEDS 2021 Challenge Page. While the entries to these challenges will focus on addressing the specific area of research, we anticipate that the dataset released as part of the challenges will also serve as a benchmark dataset for future research in eye- tracking for VR and AR.
Below is the list of topics that are of particular interest for this workshop:
- Eye-tracking in VR/AR applications
- Eye-movement analysis for biometrics, security and privacy
- Eye-tracking in medicine and health care
- Direct gaze-based interaction
- Gaze-detection techniques focusing on robustness to sensor-slippage and noise
- Novel sensors for eye tracking
- Eye-movement prediction for gaze-estimation latency compensation
- Temporal-prediction of eye-movements
- Uncertainty modeling of eye-movement prediction
- 2D/3D information fusion for semantic labeling of point-clouds
- Point-cloud segmentation
ICCV Workshop Date: Oct 17th 2021