ObjectNet3D: A Large Scale Database for 3D Object Recognition

By Yu Xiang, Wonhui Kim, Jingwei Ji, Christopher Bongsoo Choy, Hao Su, Roozbeh Mottaghi, Leonidas J. Guibas, Silvio Savarese. 2016

302 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

Strike (With) a Pose: Neural Networks Are Easily Fooled by Strange Poses of Familiar Objects (2018, background): “, ObjectNet3D [52], because most of them do not have high-quality image textures.”
Deep Level Sets: Implicit Surface Representations for 3D Shape Inference (2019, background): “Recently, large-scale datasets of 3D objects such as ShapeNet [2] and ObjectNet3D [33] have been made available which allowed the ﬁeld to make great progress.”
Learning Shape Priors for Single-View 3D Completion and Reconstruction (2018, background): “Answers are available in the footnote. researchers have made impressive progress on the these tasks [7,52,8], making use of gigantic 3D datasets [5,60,59].”
Pix3D: Dataset and Methods for Single-Image 3D Shape Modeling (2018, methodology): “Finally, we fine-tune the annotations ourselves using the GUI offered in ObjectNet3D [63].”
Physically-Based Rendering for Indoor Scene Understanding Using Convolutional Neural Networks (2016, background): “[27] introduced a 3D object-2D image database, where 3D objects are manually aligned to 2D images.”
Learning 3D Object Categories by Looking Around Them (2017, neutral): “The typical approach to learn 3D objects is to make use of large collections of high quality CAD models such as [5] or [42], which can be used to fully supervise models to recognize the objects’ viewpoint and 3D shape.”
Rethinking Reprojection: Closing the Loop for Pose-Aware Shape Reconstruction from a Single Image (2017, background): “Some efforts have been undertaken recently to rectify this including the construction of PASCAL3D+ [27], ObjectNet3D [26] and IKEA [13] datasets.”
T-LESS: An RGB-D Dataset for 6D Pose Estimation of Texture-Less Objects (2017, background): “The PASCAL3D+ [57] and the ObjectNet3D [56] provide generic 3D models and ground truth 6D poses, but only RGB images.”
StarMap for Category-Agnostic Keypoint and Viewpoint Estimation (2018, methodology): “Object-Net3D [43] contains about 50k training samples in total, but only 20k of them have keypoint annotations.”
Implicit Surface Representations As Layers in Neural Networks (2019, background): “Recently, large-scale datasets of 3D objects such as ShapeNet [2] and ObjectNet3D [34] have been made available which allowed the field to make great progress.”
ABO: Dataset and Benchmarks for Real-World 3D Object Understanding (2021, background): “Similarly, Pascal3D+ [60] and ObjectNet3D [59] provide 2D-3D alignment for images and provide more instances and categories, however the 3D annotations are only approximate matches.”
PointNetLK: Robust & Efficient Point Cloud Registration Using PointNet (2019, background): “For instance, [37, 16, 21, 44, 24], directly regress over the Euler angles of object orientations from cropped object images.”
Ordinal Depth Supervision for 3D Human Pose Estimation (2018, background): “[59, 58], humans align 3D CAD models with single images to provide viewpoint information.”
3D Bounding Box Estimation Using Deep Learning and Geometry (2016, background): “With the introduction of new challenging datasets [2, 26, 25, 12], 3D pose estimation has been extended to object categories, which requires handling both the appearance variations due to pose changes and the appearance variations within the category [9, 15].”
Review: deep learning on 3D point clouds (2020, background): “[53] has contributed a large-scale dataset for 3D object recognition.”
Weakly Supervised 3D Reconstruction with Adversarial Constraint (2017, methodology): “In Figure S3, we visualized more qualitative reconstruction results on ObjectNet3D[52] dataset.”