Generative modelling of BRDF textures from flash images
Generative modelling of BRDF textures from flash images
复制标题
DOI：
10.1145/3478513.3480507
复制
发表时间：
2021-02
期刊：
影响因子：
--
通讯作者：
P. Henzler;V. Deschaintre;N. Mitra;Tobias Ritschel
中科院分区：
文献类型：
--
作者：
P. Henzler;V. Deschaintre;N. Mitra;Tobias Ritschel
我们学习了一个潜在的空间，以便于捕获，一致的内插，并有效地再现视觉材料的外观。当用户提供在手电筒照明下拍摄的静止自然材料的照片时，首先将其转换为潜在材料代码。然后，在第二步，在材料代码的条件下，我们的方法产生无限和多样化的BRDF模型参数(漫反射反照率、法线、粗糙度、镜面反照率)的空间场，随后允许在复杂的场景和照明下进行渲染，匹配输入照片的外观。技术上，我们使用卷积编码器将所有Flash图像联合嵌入到潜在空间中，并在这些潜在代码的条件下，使用卷积神经网络(CNN)将随机空间场转换为BRDF参数场。我们调节这些BRDF参数以匹配匹配光下输入的视觉特征(视觉特征的统计和光谱)。一项用户研究将我们的方法与以前的工作进行了有利的比较，即使是那些能够获得BRDF监督的工作。项目网页：https://henzler.github.io/publication/neuralmaterial/.
We learn a latent space for easy capture, consistent interpolation, and efficient reproduction of visual material appearance. When users provide a photo of a stationary natural material captured under flashlight illumination, first it is converted into a latent material code. Then, in the second step, conditioned on the material code, our method produces an infinite and diverse spatial field of BRDF model parameters (diffuse albedo, normals, roughness, specular albedo) that subsequently allows rendering in complex scenes and illuminations, matching the appearance of the input photograph. Technically, we jointly embed all flash images into a latent space using a convolutional encoder, and -conditioned on these latent codes- convert random spatial fields into fields of BRDF parameters using a convolutional neural network (CNN). We condition these BRDF parameters to match the visual characteristics (statistics and spectra of visual features) of the input under matching light. A user study compares our approach favorably to previous work, even those with access to BRDF supervision. Project webpage: https://henzler.github.io/publication/neuralmaterial/.