120 datasets found

HomebrewedDB (HB) is a scenes dataset [23] used for 6D pose estimation that features high-quality 3D-reconstructed models for a total of 33 instances.

ARID40k is a subset of a ARID [29], containing 6000 scenes rom 153 different object instances in 51 categories.

Washington RGB-D Objects dataset (ROD) is a well-established benchmark for object recognition tasks in RGB-D domains. It contains up to 41 877 views from 300 object instances,...

The N-Caltech 101 dataset is an event-based object recognition dataset generated from the Caltech 101 object recognition dataset.

N-ImageNet is a large-scale dataset containing 1,781,167 event stream data of 1,000 object classes.

The SHREC dataset contains 3D models of various objects, including furniture, vehicles, and building components.

ModelNet is a large-scale 3D object recognition dataset containing 30,000 models from 50 categories.

The Objaverse dataset contains around 800k 3D objects. After adopting simple filter leveraging CLIP [27] to remove the objects whose rendered images are not relevant to its...

VisDA-2017 is a simulation-to-real dataset with two extremely distinct domains: Synthetic renderings of 3D models and Real collected from photo-realistic or real-image datasets.

The dataset used in this paper is a neural network, and the authors used it to test the performance of their lookahead pruning method.

CORe50 is a video benchmark for continual learning, consisting of 164,866 128x128 images of 50 domestic objects belonging to 10 categories.

The dataset used for training and validation of the proposed approach to combine semantic segmentation and dense outlier detection.

The Surface Reconstruction Benchmarks (SRB) dataset contains 3D models of various objects.

The 3D Scene dataset contains real-world scenes with multiple views.

A benchmark for low-shot recognition was proposed by Hariharan & Girshick (2017) and consists of a representation learning phase without access to the low-shot classes and a...

Common Objects in 3D: Large-scale learning and evaluation of real-life 3D category reconstruction

Google Scanned Objects is a real-scanned 3D object dataset and we use all its 1030 samples for evaluation unlike existing works [29, 31] that only use 30 of them.

A large-scale dataset of 3D objects for training and testing 3D reconstruction models.

The dataset used in the paper is not explicitly described, but it is mentioned that the authors used it for language modeling and image classification tasks.

Point cloud registration is a crucial problem in computer vision and robotics. Existing methods either rely on matching local geometric features, which are sensitive to the pose...