首页> 外文期刊>ACM Transactions on Graphics >Appearance-from-Motion: Recovering Spatially Varying Surface Reflectance under Unknown Lighting
【24h】

Appearance-from-Motion: Recovering Spatially Varying Surface Reflectance under Unknown Lighting

机译:运动外观:在未知光照下恢复空间变化的表面反射率

获取原文
获取原文并翻译 | 示例
           

摘要

We present “appearance-from-motion”, a novel method for recoveringrnthe spatially varying isotropic surface reflectance from a videornof a rotating subject, with known geometry, under unknown naturalrnillumination. We formulate the appearance recovery as an iterativernprocess that alternates between estimating surface reflectancernand estimating incident lighting. We characterize the surface reflectancernby a data-driven microfacet model, and recover the microfacetrnnormal distribution for each surface point separately fromrntemporal changes in the observed radiance. To regularize the recoveryrnof the incident lighting, we rely on the observation that naturalrnlighting is sparse in the gradient domain. Furthermore, we exploitrnthe sparsity of strong edges in the incident lighting to improve thernrobustness of the surface reflectance estimation. We demonstraternrobust recovery of spatially varying isotropic reflectance from capturedrnvideo as well as an internet video sequence for a wide varietyrnof materials and natural lighting conditions.
机译:我们提出“运动出现”的新方法,该方法可以在未知的自然光照下,从已知几何形状的旋转物体中恢复视频物体在空间上各向同性的表面反射率。我们将外观恢复公式化为一个迭代过程,该过程在估计表面反射率和估计入射光之间交替进行。我们通过数据驱动的微面模型表征表面反射率,并从观察到的辐射度的时空变化中分别恢复每个表面点的微面正态分布。为了规范入射光的恢复,我们依靠观察到自然光在梯度域中是稀疏的。此外,我们利用入射光中强边缘的稀疏性来提高表面反射率估计的稳健性。我们展示了从捕获的视频以及各种视频和自然采光条件下的互联网视频序列中对空间变化各向同性反射的强健恢复。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号