Update:
Currently there seems to be a preference for using point clouds over raw meshes for 3d segmentation. They can use all the benchmarks below plus:
- S3DIS
URL: http://buildingparser.stanford.edu/dataset.html
- ScanNet
URL: http://www.scan-net.org/
The state of the art seems to be [11] closely followed by [10] and both use PointNet++ [12].
The one you are refering to is the common PSB (Princeton Segmentation Benchmark) dataset [1].
Other Important datasets used for 3D mesh segmentation are:
- COSEG [2]
- ShapeNet [3]
- Other from Princeton [4]
- Other basen on subset of ShapeNet [5]
On the other hand, there seems to be no state-of-the-art page for 3D shape segmentation. But the papers I find most important are [6, 7, 8, 9], with 9 the state of the art.
Refs:
- Chen et al: A Benchmark for 3D Mesh Segmentation, 2009
- Wang et al:
Active co-analysis of a set of shapes, 2012
- Kim et al: Learning Part-based Templates from Large Collections of 3D Shapes, 2013
- Chang et al: ShapeNet: An Information-Rich 3D Model Repository, 2015
- Yi et al: A scalable active framework for region annotation in 3d shape collections, 2016
- Kalogerakis et al: Learning 3D Mesh Segmentation and Labeling, 2010
- Guo et al: 3D Mesh Labeling via Deep Convolutional Neural Networks, 2015
- Kalogerakis et al: 3D Shape Segmentation with Projective Convolutional Networks, 2017
- Wang et al: 3D shape segmentation via shape fully convolutional networks, 2018
- Wang et al: Associatively Segmenting Instances and Semantics in Point Clouds, 2019
- Yang et al: Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds, 2019
- Qi et al: PointNet++: Deep Hierarchical Feature Learning on
Point Sets in a Metric Space, 2017