Foveated rendering techniques have seen recent development with the advent of commercial head-mounted displays with eye-tracking capabilities. The main drive is to exploit particular features of our peripheral vision ...
详细信息
Foveated rendering techniques have seen recent development with the advent of commercial head-mounted displays with eye-tracking capabilities. The main drive is to exploit particular features of our peripheral vision that allow optimizing rendering pipelines, which allows using less computational effort where the human visual system may be unaware of differences. Most efforts have been focused on simplifying spatial visual detail on areas not being focused on by adjusting acuity of shading models, sharpness of images, and pixel density. However, other perception pipeline areas are also influential, particularly in certain purpose-specific applications. In this paper, we demonstrate it is possible to reduce animation rates in crowd simulations up to a complete stop for agents in our peripheral vision without users noticing the effect. We implemented a prototype Unity3D application with typical crowd simulation scenarios and carried out user experiments to study subjects’ perception to changes in animation rates. We find that in the best case we were able to reduce the number of operations by 99.3% compared to an unfoveated scenario, with opportunities for developments combined with other acceleration techniques. This paper also includes an in-depth discussion about human perception of movement in peripheral vision with novel ideas that will have applications beyond crowd simulation.
Gaze-tracking has a wide range of applications across scientific and industrial fields, and recent computer vision and deep-learning advances have made gaze-tracking with standard webcams possible. However, current so...
详细信息
Gaze-tracking has a wide range of applications across scientific and industrial fields, and recent computer vision and deep-learning advances have made gaze-tracking with standard webcams possible. However, current solutions only offer suboptimal performance and lack flexibility. This paper introduces GazeFollower, an accessible system for webcam gaze-tracking in Python. GazeFollower stands out for its customizability, allowing researchers to quickly develop and adapt algorithms to meet their needs. At its core, GazeFollower estimates gaze with a model trained on 32 million face images, ensuring robust gaze tracking. A benchmark test on a sizeable sample (N=31) shows that the tracking performance of GazeFollower is on par with or better than budget commercial eye trackers. With calibration, GazeFollower has an accuracy of 1.11 cm and a precision of 0.11 cm, and personalized model fine-tuning further enhances these metrics to 0.92 cm and 0.08 cm. These results suggest that GazeFollower holds potential for real-world applications.
The adaptation and generaliza-{Ton-6r-scan and slit-scan animatioh stand techniques for use in computer generated animation is discussed. Scan and slitscan techniques are based on moving artwork, camera, and, for slit...
详细信息
ISBN:
(纸本)0897910214
The adaptation and generaliza-{Ton-6r-scan and slit-scan animatioh stand techniques for use in computer generated animation is discussed. Scan and slitscan techniques are based on moving artwork, camera, and, for slit-scan, a thin aperture while the camera shutter is open. These processes can be described as selectively sampling an environment over time and recording the result as a single image. Sequences of such images form the animated film. Rather than use mechanical means to accomplish this, it is possible to develop algorithms which mimic this processes but are based on sampling dynamic environment descriptions to generate computer produced images. The use of computergraphics allows these techniques to be generalized in ways difficult or impossible even with very elaborate animation stands. Use of multiple independent scanning apertures and three-dimensional environments are natural generalizations. The exact algorithms used depend on the characteristics of the graphics systems used. An approach based on using a real-time shaded graphics system and an approach using frame buffer systems are outlined. The first approach can also be applied to refresh vector graphics systems. � acm. All rights reserved.
暂无评论