Geometric reasoning in computer vision with using only 2D supervision

仅使用 2D 监督的计算机视觉中的几何推理

• 批准号: DP220103803
• 负责人: Prof Simon Lucey
• 金额: $ 31.33万
• 依托单位: The University of Adelaide
• 依托单位国家: 澳大利亚
• 项目类别: Discovery Projects
• 财政年份: 2022
• 资助国家: 澳大利亚
• 起止时间: 2022-01-01 至 2024-12-31
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/DP220103803
• 关键词: Geometric; reasoning; computer; vision; using

The aim of the project is to build a geometric reasoning system that can exhibit human like performance. Advances in autonomous systems such as vehicles, robots, and drones will transform the Australian and global economy. Geometric reasoning is fundamental to advancement in such AI and is the focus of this project. The project will leverage a theoretical breakthrough in the field of structure from motion; which will allow an AI to learn the 3D pose and shape of an object solely through 2D supervision. The project will provide new insights into how AI should understand the 3D world.

该项目的目的是建立一个几何推理系统，可以表现出人类一样的性能。汽车、机器人和无人机等自主系统的进步将改变澳大利亚和全球经济。几何推理是这种人工智能进步的基础，也是这个项目的重点。该项目将利用运动结构领域的理论突破;这将允许人工智能仅通过2D监督来学习物体的3D姿态和形状。该项目将为人工智能如何理解3D世界提供新的见解。