992 datasets found

Semantic segmentation is a crucial and challenging task for image understanding. It aims to predict a dense labeling map for the input image, which assigns each pixel a unique...

The dataset used for 6D object pose estimation, consisting of images of 16 objects with varying levels of occlusion.

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...

The datasets used in the paper are described as follows: WikiMIA, BookMIA, Temporal Wiki, Temporal arXiv, ArXiv-1 month, Multi-Webdata, LAION-MI, Gutenberg.

Recent studies show that Transformer has strong capability of building long-range dependencies, yet is incompetent in capturing high frequencies that predomi- nantly convey...

Self-supervised representation learning for Vision Transformers (ViT) to mitigate the greedy needs of ViT networks for very large fully-annotated datasets.

The dataset used in the paper is ShapeNetPart, a synthetic dataset for 3D object part segmentation. It contains 16,881 models from 16 categories.

Large scale multi-view stereopsis evaluation dataset.

The dataset used in this study is a semantic PCB image dataset, which contains images of PCBs with annotated components.

The dataset used in the paper is not explicitly described, but it is mentioned that the authors conducted experiments on various vision tasks, including image classification,...

An N-object dataset for testing the proposed framework

The dataset used in the paper SpikingNeRF: Making Bio-inspired Neural Networks See through the Real World

The dataset used in the paper SpikingNeRF: Making Bio-inspired Neural Networks See through the Real World

A dataset of 650 machine learning models presented in academic publications and relevant gray literature.

Multiscale Vision Transformers (MViT) for video and image recognition, by connecting the seminal idea of multiscale feature hierarchies with transformer models.

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,...

RGB-D scene recognition approaches often train two standalone backbones for RGB and depth modalities with the same Places or ImageNet pre-training. However, the pre-trained...

We propose DisCo-CLIP, a distributed memory-efficient CLIP training approach, to reduce the memory consump- tion of contrastive loss when training contrastive learning models.

Object recognition is arguably the most important problem at the heart of computer vision. Recently, Barbu et al. introduced a dataset called ObjectNet which includes objects in...

Progress in GANs has enabled the generation of high-res-olution photorealistic images of astonishing quality. StyleGANs allow for compelling attribute modification on such...