arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1808.10232cs.CV
收藏

用于训练和验证的场景流数据自动生成

Automated Scene Flow Data Generation for Training and Verification

DFKI(德国人工智能研究中心);Bit-TS(Bit-TS公司);BMW(宝马集团)

Oliver Wasenmüller, René Schuster, Didier Stricker, Karl Leiss, Jürger Pfister, Oleksandra Ganus, Julian Tatsch, Artem Savkin, Nikolas Brasch

摘要:

场景流描述了图像中每个像素的3D位置以及3D运动。这些算法是许多最先进的自主或自动驾驶功能的基础。为了验证和训练,需要大量的地面实况数据,这对于真实数据是不可用的。在本文中,我们展示了一种技术,以创建合成数据与密集和精确的场景流地面实况。

英文摘要:

Scene flow describes the 3D position as well as the 3D motion of each pixel in an image. Such algorithms are the basis for many state-of-the-art autonomous or automated driving functions. For verification and training large amounts of ground truth data is required, which is not available for real data. In this paper, we demonstrate a technology to create synthetic data with dense and precise scene flow ground truth.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑