The current production model for 3D scenes, inherited from the early years of computer graphics, prevents a real collaboration among the subjects involved and, as a consequence, the full exploitation of the new potent...
详细信息
ISBN:
(纸本)0780371984
The current production model for 3D scenes, inherited from the early years of computer graphics, prevents a real collaboration among the subjects involved and, as a consequence, the full exploitation of the new potentialities of the 3D media. In this scenario the content expert is only a consultant of the 3d modeler, who is directly responsible for designing the interaction model. In complex 3D applications interactivity, graphic complexity and content play different important roles that should correspond to different well defined skills: authoring is a collaborative activity that involves different experts. The content expert is the key figure of the extended team of authors that collaborate to the design and implementation of a complex interactive application. He/she is an intermediate actor between the final users and the experts devoted to the design of the low level graphic details: is an expert in the application domain and understands the needs of the final users with respect to orientation, navigation and access to the information content; is able to speak the technical language needed to describe the interaction but needs to be provided with high level interaction interface, built by the graphic experts, because is not necessarily skilled in the fine-grained language of interaction with the technology used to build virtual worlds.
The development of specific algorithms in image processing are usually related to dataset characteristics. Those characteristics will influence the number of instructions required to solve a problem. Normally, the mor...
详细信息
This article presents ArCo which is an architecture for managing a digital environment in a context of ambient intelligence. ArCo can integrate all kinds of objects and is totally configurable. Moreover it proposes a ...
详细信息
This article presents ArCo which is an architecture for managing a digital environment in a context of ambient intelligence. ArCo can integrate all kinds of objects and is totally configurable. Moreover it proposes a visual programming interface, AmbiProg, which allows controlling each digital object which composes an environment. The objective of this paper is to appraise AmbiProg. An experiment was realized with 16 children around 10 years old. They were asked to realize exercises in a limited period of time. The objective was to determine if programming was easy and how many times they needed to learn scenarios creation. It appeared that they succeed tasks and were able to use AmbiProg. Programming an environment was easy as pie.
The recent advances in deep neural networks have led to effective vision-based reinforcement learning methods that have been employed to obtain human-level controllers in Atari 2600 games from pixel data. Atari 2600 g...
详细信息
The recent advances in deep neural networks have led to effective vision-based reinforcement learning methods that have been employed to obtain human-level controllers in Atari 2600 games from pixel data. Atari 2600 games, however, do not resemble real-world tasks since they involve non-realistic 2D environments and the third-person perspective. Here, we propose a novel test-bed platform for reinforcement learning research from raw visual information which employs the first-person perspective in a semi-realistic 3D world. The software, called ViZDoom, is based on the classical first-person shooter video game, Doom. It allows developing bots that play the game using the screen buffer. ViZDoom is lightweight, fast, and highly customizable via a convenient mechanism of user scenarios. In the experimental part, we test the environment by trying to learn bots for two scenarios: a basic move-and-shoot task and a more complex maze-navigation problem. Using convolutional deep neural networks with Q-learning and experience replay, for both scenarios, we were able to train competent bots, which exhibit human-like behaviors. The results confirm the utility of ViZDoom as an AI research platform and imply that visual reinforcement learning in 3D realistic first-person perspective environments is feasible.
暂无评论