Abstract: Systems and methods are disclosed for rendering a surface in a synthetic scene using gutter space padding for texture atlases. Gutter space adjacent to UV shells may be padded with texels from the UV shells used for neighboring surfaces, rather than merely extending the interior texels. The result is that distracting artifacts, caused by mesh downsampling that produces sub-optimal polygon shapes and vertex placement may be mitigated. The resulting rendering, manifesting lessened artifacts, may appear to be less impacted by downsampling, and thus improve the experience for the viewer or other synthetics data consumer. The improvement may be used in rendering for human users and also training neural networks in computer vision tasks, such as object detection and recognition.

Abstract: Systems and methods are disclosed for rendering a surface in a synthetic scene using gutter space padding for texture atlases. Gutter space adjacent to UV shells may be padded with texels from the UV shells used for neighboring surfaces, rather than merely extending the interior texels. The result is that distracting artifacts, caused by mesh downsampling that produces sub-optimal polygon shapes and vertex placement may be mitigated. The resulting rendering, manifesting lessened artifacts, may appear to be less impacted by downsampling, and thus improve the experience for the viewer or other synthetics data consumer. The improvement may be used in rendering for human users and also training neural networks in computer vision tasks, such as object detection and recognition.

Abstract: A system for training a neural network generates a plurality of training meshes based on an input mesh. The plurality of training meshes include at least one mesh perceptually similar to the input mesh and one arbitrarily selected mesh perceptually dissimilar to the input mesh. The neural network is trained using the input mesh and the plurality of training meshes by tuning output of the neural network to identify similar non-identical meshes. The trained neural network is robust in identifying meshes similar to an unknown mesh input to the trained neural network.