DiffusionNet: Discretization Agnostic Learning on Surfaces
نویسندگان
چکیده
We introduce a new general-purpose approach to deep learning on three-dimensional surfaces based the insight that simple diffusion layer is highly effective for spatial communication. The resulting networks are automatically robust changes in resolution and sampling of surface—a basic property crucial practical applications. Our can be discretized various geometric representations, such as triangle meshes or point clouds, even trained one representation then applied another. optimize support continuous network parameter ranging from purely local totally global, removing burden manually choosing neighborhood sizes. only other ingredients method multi-layer perceptron independently at each gradient features directional filters. simple, robust, efficient. Here, we focus primarily mesh demonstrate state-of-the-art results variety tasks, including surface classification, segmentation, non-rigid correspondence.
منابع مشابه
Agnostic G1 Gregory surfaces
We discuss G smoothness conditions for rectangular and triangular Gregory patches. We then incorporate these G conditions into a surface fitting algorithm. Knowledge of the patch type is inconsequential to the formulation of the G conditions, hence the term agnostic G Gregory surfaces. 2012 Elsevier Inc. All rights reserved.
متن کاملAgnostic G Gregory Surfaces
We discuss G smoothness conditions for rectangular and triangular Gregory patches. We then incorporate these G conditions into a surface fitting algorithm. Knowledge of the patch type is inconsequential to the formulation of the G conditions, hence the term agnostic G Gregory surfaces.
متن کاملAgnostic Online Learning
We study learnability of hypotheses classes in agnostic online prediction models. The analogous question in the PAC learning model [Valiant, 1984] was addressed by Haussler [1992] and others, who showed that the VC dimension characterization of the sample complexity of learnability extends to the agnostic (or ”unrealizable”) setting. In his influential work, Littlestone [1988] described a combi...
متن کاملToward Eecient Agnostic Learning
In this paper we initiate an investigation of generalizations of the Probably Approximately Correct (PAC) learning model that attempt to signiicantly weaken the target function assumptions. The ultimate goal in this direction is informally termed agnostic learning, in which we make virtually no assumptions on the target function. The name derives from the fact that as designers of learning algo...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: ACM Transactions on Graphics
سال: 2022
ISSN: ['0730-0301', '1557-7368']
DOI: https://doi.org/10.1145/3507905