Appearance-Driven Automatic 3D Model Simplification

  title={Appearance-Driven Automatic 3D Model Simplification},
  author={J. Hasselgren and J. Munkberg and J. Lehtinen and M. Aittala and S. Laine},
Fig. 1. We automatically approximate shape and appearance of detailed 3D scenes. Here, we approximate foliage with low-poly proxy geometry and textures, instance it 3000 times, and render the scene in a standalone path tracer. Compared to the reference, we accurately capture the appearance of the scene at just a fraction (0.4%) of the triangle count. The geometry is illustrated in the insets. The assets are from the Moana Island Scene, a publicly available data set courtesy of Walt Disney… Expand


Image-driven simplification
We introduce the notion of image-driven simplification, a framework that uses images to decide which portions of a model to simplify. This is a departure from approaches that make polygonalExpand
Stochastic simplification of aggregate detail
A stochastic technique related to some approaches used for point-based renderers is introduced, which aims to address the problem of poor performance on scenes that contain a lot of detailed geometry. Expand
Billboard clouds for extreme model simplification
This work introduces billboard clouds -- a new approach for extreme simplification in the context of real-time rendering that combines the strengths of mesh decimation and image-based impostors. Expand
Image-based reconstruction of spatial appearance and geometric detail
This work presents an image-based measuring method that robustly detects the different materials of real objects and fits an average bidirectional reflectance distribution function (BRDF) to each of them, leading to a truly spatially varying BRDF representation. Expand
Learning to Predict 3D Objects with an Interpolation-based Differentiable Renderer
A differentiable rendering framework which allows gradients to be analytically computed for all pixels in an image and to view foreground rasterization as a weighted interpolation of local properties and background rasterized as a distance-based aggregation of global geometry. Expand
Surface simplification using quadric error metrics
This work has developed a surface simplification algorithm which can rapidly produce high quality approximations of polygonal models, and which also supports non-manifold surface models. Expand
Modular primitives for high-performance differentiable rendering
A modular differentiable renderer design that yields performance superior to previous methods by leveraging existing, highly optimized hardware graphics pipelines, and allows custom, high-performance graphics pipelines to be built directly within automatic differentiation frameworks such as PyTorch or TensorFlow. Expand
Model simplification using vertex-clustering
The technique is advanced with careful consideration ofroximation quality and smoothness in transitions between different levels of simplification, while maintaining its efficiency and efb%iveness. Expand
View-dependent simplification of arbitrary polygonal environments
HDS is dynamic, retessellating the scene continually as the user's viewing position shifts, and global, processing the entire database without first decomposing the environment into individual objects. Expand
Skinning mesh animations
An automatic algorithm for generating progressive skinning approximations, that is particularly efficient for pseudo-articulated motions, is provided, which includes the use of nonparametric mean shift clustering of high-dimensional mesh rotation sequences to automatically identify statistically relevant bones, and robust least squares methods to determine bone transformations. Expand