A benchmark for 3D mesh segmentation

By Xiaobai Chen, Aleksey Golovinskiy, Thomas A. Funkhouser. 2009

786 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

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 Vision and Pattern Recognition

Citations in the corpus

ShapeNet: An Information-Rich 3D Model Repository (2015, methodology): “For example, the Benchmark for 3D Mesh Segmentation contains just 380 models in 19 object classes [2].”
Efficient Decomposition of Image and Mesh Graphs by Lifted Multicuts (2015, methodology): “We have assessed both algorithms in conjunction with both optimization problems in applications to image decomposition (BSDS-500 benchmark [8]) and mesh decomposition (Princeton benchmark [16]).”
Deepshape: Deep learned shape descriptor for 3D shape matching and retrieval (2015, background): “Nowadays there is an explosive growth of 3D meshed surface models in a variety of fields, such as engineering, entertainment and medical imaging [23, 20, 16, 10, 9, 6].”
3D deep shape descriptor (2015, background): “With recent advancements in 3D acquisition and printing techniques, we have observed an exponential increase in 3D-meshed surface models across a variety of fields, such as engineering, entertainment, and medical imaging [41, 36, 28, 12, 11, 8, 40, 1].”
MaskFusion: Real-Time Recognition, Tracking and Reconstruction of Multiple Moving Objects (2018, background): “Scene segmentation: The computer graphics [7] and vision [9, 17, 22,28,46] communities have devoted substantial effort to object and scene segmentation.”
Registration of 3D Point Clouds and Meshes: A Survey from Rigid to Nonrigid (2013, neutral): “benchmark for three-dimensional mesh segmentation [110] (e.”
Beyond Point Clouds: Scene Understanding by Reasoning Geometry and Physics (2013, background): “Two evaluation criterion: “Cut Discrepancy” and “Hamming Distance” mentioned in [5] are adopted.”
Learning 3D Part Detection from Sparsely Labeled Data (2014, background): “A portion of the Princeton shape dataset [21] (PSB) has been densely labeled by [2].”
Joint shape segmentation with linear programming (2011, methodology): “2008]) and randomized cuts (RC, [Golovinskiy and Funkhouser 2008]); statistics for the other algorithms are provided in the original benchmark [Chen et al. 2009].”
3D Mesh Labeling via Deep Convolutional Neural Networks (2015, methodology): “These experiments are conducted on the Princeton Segmentation Benchmark (PSB) [Chen et al. 2009] (19 categories, 20 meshes per category) and 4 additional categories including Lamp (20 meshes), Candelabra (28 meshes), Iron (18 meshes) and Guitar (44 meshes) from van Kaick et al. [2011], Fig.”
3D Shape Segmentation and Labeling via Extreme Learning Machine (2014, neutral): “Huang et al. [HKG11] presented a novel linear programming approach to joint segmentation of a heterogeneous shape collection, producing comparable results to the supervised approaches on the Princeton Segmentation Benchmark (PSB) [CGF09].”
A Review on Deep Learning Techniques Applied to Semantic Segmentation (2017, neutral): “• A Benchmark for 3D Mesh Segmentation [52]21: this benchmark is composed by 380 meshes classified in 19 categories (human, cup, glasses, airplane, ant, chair, octopus, table, teddy, hand, plier, fish, bird, armadillo, bust, mech, bearing, vase, fourleg).”