PartNet: A Large-Scale Benchmark for Fine-Grained and Hierarchical Part-Level 3D Object Understanding

By Kaichun Mo, Shilin Zhu, Angel X. Chang, Li Yi, Subarna Tripathi, Leonidas J. Guibas, Hao Su. 2018

953 citations recorded by the source corpus.

We present PartNet: a consistent, large-scale dataset of 3D objects annotated with fine-grained, instance-level, and hierarchical 3D part information. Our dataset consists of 573,585 part instances over 26,671 3D models covering 24 object categories. This dataset enables and serves as a catalyst for many tasks such as shape analysis, dynamic 3D scene modeling and simulation, affordance analysis, and others. Using our dataset, we establish three benchmarking tasks for evaluating 3D part recognition: fine-grained semantic segmentation, hierarchical semantic segmentation, and instance segmentation. We benchmark four state-of-the-art 3D deep learning algorithms for fine-grained semantic segmentation and three baseline methods for hierarchical semantic segmentation. We also propose a baseline method for part instance segmentation and demonstrate its superior performance over existing methods.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, 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 Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

Citations in the corpus

A Survey of Embodied AI: From Simulators to Research Tasks (2021, methodology): “The clear object segmentation for the 3D assets makes it easy to model them as articulated objects with movable joints, such as the 3D models provided in PartNet [36].”
SAPIEN: A SimulAted Part-Based Interactive ENvironment (2020, methodology): “To address the lack-of-content problem, our work provides a large-scale simulation-ready dataset, PartNet-Mobility dataset, that is constructed from 3D model datasets including PartNet [33] and ShapeNet [7].”
PointOdyssey: A Large-Scale Synthetic Dataset for Long-Term Point Tracking (2023, methodology): “Similar to Kubric [24], we populate out-door scenes with random rigid objects from GSO [19] and PartNet [42].”
OpenRooms: An Open Framework for Photorealistic Indoor Scene Datasets (2020, methodology): “While Photoshape does this for only chairs, we do so for all furniture types in indoor scenes, using the semantically meaningful part segmentation of 24 categories of models provided by PartNet [41].”
iGibson, a Simulation Environment for Interactive Tasks in Large Realistic Scenes (2020, methodology): “The object models are curated from open-source datasets: ShapeNet [21], PartNet Mobility [15, 43], and SketchFab.”
OpenRooms: An End-to-End Open Framework for Photorealistic Indoor Scene Datasets (2020, methodology): “Thus, we use the segmentation from PartNet [34], which provides high-quality semantically meaningful part segmentation of 24 categories of ShapeNet models.”
PTR: A Benchmark for Part-based Conceptual, Relational, and Physical Reasoning (2021, methodology): “The PTR universe contains ﬁve categories of objects (chair, table, bed, refrigerator and cart) from the PartNet dataset.”
Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts (2020, background): “Recent advances in deep learning on point clouds, such as those obtained from LiDAR or depth sensors, together with a proliferation of public, annotated datasets [9, 13, 53, 32, 65, 2, 40, 56], have led to swift progress in 3D scene understanding.”
GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud (2018, background): “[27] introduces PartNet, a large scale fine-grained part annotation dataset, where 3D objects from [3] are segmented into fine-grained part instances.”
3D Shape Generation and Completion through Point-Voxel Diffusion (2021, methodology): “Different from PartNet, ShapeNet’s partial shapes are sampled from customrendered depth images.”
HOI4D: A 4D Egocentric Dataset for Category-Level Human-Object Interaction (2022, methodology): “For articulated objects, we additionally provide part annotations similar to PartNet [34].”
Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds (2019, methodology): “formulate the instance segmentation as a per-point feature classification problem in PartNet [32].”
RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation (2025, positive): “SAPIEN [40] supports fully dynamic interaction with over 2,300 articulated objects from PartNet-Mobility [40, 28], ManiSkill2 supplies more than four million demonstration frames across twenty task families [14], Meta-World [42] offers fifty standardized manipulation tasks for multi-task and…”
ReferIt3D: Neural Listeners for Fine-Grained 3D Object Identification in Real-World Scenes (2020, methodology): “For (a) we utilized the Scan2CAD [14] annotations that provide 9DOF alignments between ShapeNet models and ScanNet objects, and for (b) we used a combination of PartNet’s [45] and manual annotations.”