18 datasets found

The CAD dataset is a photorealistic 3D generation dataset conditioned on a single image and a text prompt.

The dataset used in the paper is a MPI3D dataset, which contains 3D images of objects with varying sizes and colors.

The dataset used for evaluating latent-space interpolation algorithms for variational autoencoders.

The dataset used in the paper is ShapeNet, a large-scale 3D shape dataset, and ModelNet40, a dataset for 3D object classification.

The ModelNet10 dataset contains 8,000 training and 2,000 test 3D models of objects.

The dataset used in the paper Strike (with) a Pose: Neural Networks Are Easily Fooled by Strange Poses of Familiar Objects. The dataset consists of 30 unique 3D object models...

The Deitke dataset contains 30 objects chosen by SyncDreamer [17] and rendered 16 views with uniformly distributed camera poses and environment lighting for each object.

The 3DTopia-360K dataset is a large-scale 3D object dataset, which is used to train the 3DTopia model. The dataset contains 360K 3D objects with detailed captions.

The dataset used in the Zero-1-to-3 paper, containing 10M 3D objects.

The PartNet dataset is a large-scale benchmark for 3D object recognition and retrieval. It consists of 527 objects composed of 2690 parts, which can be composited into a...

The dataset used in the paper for text-to-3D generation, consisting of 5,000 different prompts and 800,000 generated images.

Zero-shot learning (ZSL) aims to classify objects that are not observed or seen during training. It relies on class semantic description to transfer knowledge from the seen...

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...

CLEVR images contain objects characterized by a set of attributes (shape, color, size and material). The questions are grouped into 5 categories: Exist, Count, CompareInteger,...

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.

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...