Learning part-based templates from large collections of 3D shapes

By Vladimir G. Kim, Wilmot Li, Niloy J. Mitra, Siddhartha Chaudhuri, Stephen J. DiVerdi, Thomas A. Funkhouser. 2013

253 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning, 3D Vision & Pose

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

In 3D Vision & Pose

Citations in the corpus

SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation (2016, methodology): “For the keypoint prediction task, we build a large scale keypoint annotation dataset, containing 1,337 chair models with 10 keypoints per shape, in contrast to traditional small scale dataset [14] which has at most 100 shapes annotated per category.”
ShapeNet: An Information-Rich 3D Model Repository (2015, neutral): “Other commonly-used datasets contain segmentations [2], correspondences [13, 12], hierarchies [19], symmetries [11], salient features [3], semantic segmentations and labels [36], alignments of 3D models with images [35], semantic ontologies [5], and other functional annotations — but again only for…”
Lifting Object Detection Datasets into 3D (2015, methodology): “extend the domain of powerful shape analysis techniques developed in the graphics community [62], that have however been mostly applied to collections of CAD models.”
A Field Model for Repairing 3D Shapes (2016, background): “Specifically, the 3D surfaces are missing and/or broken and this phenomenon causes difficulties for many sequential tasks such as 3D object detection and recognition [30, 36], shape analysis [20, 19], and scene understanding [32, 12].”
PTR: A Benchmark for Part-based Conceptual, Relational, and Physical Reasoning (2021, background): “Restricted by the annotations in existing large-scale datasets, previous part-based 3D understanding works mainly focus on identifying parts [15, 30, 26, 55, 32, 58] and part relations [6, 10, 31, 35, 60, 50] for individual 3D objects.”
3D Shape Segmentation with Projective Convolutional Networks (2016, methodology): “Other approaches segment shapes by employing non-rigid alignment steps through deformable part templates [27, 20], or transfer labels through surface correspondences and functional maps between 3D shapes [48, 22, 50, 27, 23].”
PCN: Point Completion Network (2018, background): “Some [15, 24, 32, 34, 43] retrieve the complete shape directly while some [17, 19, 28, 45, 52] retrieve object parts and then assemble them to obtain the complete shape.”
Data-driven structural priors for shape completion (2015, methodology): “The key advantage of our representation over methods that use global reference frame (e.g., [Kim et al. 2013a; Shen et al. 2012]) is that any visible part can serve as a reference to guide the prediction of occluded parts.”
3D Mesh Labeling via Deep Convolutional Neural Networks (2015, methodology): “And from Table V, we can see that the proposed approach is superior to state-of-the-art techniques [Sidi et al. 2011; Kim et al. 2013] on these two datasets.”
Deformation-driven shape correspondence via shape recognition (2017, methodology): “Parallel works explored the use of part-based Bayesian networks [Chaudhuri et al. 2011; Kalogerakis et al. 2012] and modular templates [Kim et al. 2013; Fish et al. 2014] to represent both continuous and discrete variations.”