65 datasets found

A dataset for the study of domain adaptation solutions, composed of synthetic and real videos from seven gesture classes, from both the ground and air perspectives.

Stargan v2 dataset

The Hyper-Kvasir→Piccolo task is a domain adaptation task for semantic segmentation, where the source domain is Hyper-Kvasir and the target domain is Piccolo.

The Synthia→Cityscapes task is a domain adaptation task for semantic segmentation, where the source domain is Synthia and the target domain is Cityscapes.

The Office-Home dataset is a visual domain adaptation task, where the goal is to adapt a model trained in a source domain to a target domain with different distributions.

The DomainBed dataset is a benchmark for domain generalization, containing five multi-domain datasets (VLCS, PACS, OfficeHome, TerraIncognita, and DomainNet). The dataset is...

The dataset used in the paper is a large collection of small images, each representing a patch of a jigsaw puzzle. The patches are of the same size and orientation, and the goal...

The 150-class dataset is a balanced dataset with 150 intents that can be grouped into 10 general domains.

The 73-domain dataset is an imbalanced dataset with 73 domains. Each domain contains 512 examples on average.

The dataset used for out-of-distribution segmentation, zero-shot semantic segmentation, and domain adaptation.

Biwi Kinect is a real-world dataset containing over 15K images of 20 people.

MPI3D is a benchmark dataset that consists of 1,036,800 examples of 3D objects from three different domains: Toy, RealisticC, and Real.

The dataset is used for deep hashing network for unsupervised domain adaptation.

Test time adaptation (TTA) aims to adapt deep neural networks when receiving out of distribution test domain samples.

A popular dataset for computer vision, where the first domain contains the original images, and the second is generated by applying brightness, gray scaling, and Gaussian blurring.

Jointly measured observations of single-cell RNA-seq and ATAC-seq data.

One domain consists of the original MNIST digits, while the other is generated by applying multiple transformations including rotation, downscaling, and Gaussian blurring.

The data collection process of a given phenomenon may be affected by different sources of variability, creating seemingly distinct domains.

Multi-domain data is becoming increasingly common and presents both challenges and opportunities in the data science community. The integration of distinct data-views can be...

Wearable Human Activity Recognition (WHAR) models usually face performance degradation on the new user due to user variance. Unsupervised domain adaptation (UDA) becomes the...