57 datasets found

The dataset used for local anomaly detection in videos using object-centric adversarial learning.

A video anomaly detection dataset for a crowded scene, containing 11 short clips of people meandering around an outdoor field, courtyard, or indoor foyer.

UCF101, HMDB51, Olympic Sports Dataset

The UCF101 dataset and videos captured using mobile phones and webcam representing the activity were used for training, inference, reconfiguration, and evaluation of the hardware.

RealityTalk is a system that augments live video presentations with speech-driven interactive virtual elements.

The dataset used in the paper for few-shot video classification, containing videos from Kinetics and Something-Something V2 datasets.

The EDU dataset involves hand-crafted features extracted for 5,611 students across 10 weeks of data.

The HOT-2023 dataset is a hyperspectral object tracking dataset, which consists of 110 training videos and 87 validation videos. The dataset includes 16, 15, and 25 bands for...

The HOT-2022 dataset is a hyperspectral object tracking dataset, which consists of 40 training videos and 35 testing videos. The dataset includes 16 bands and has a resolution...

A dataset of eye-tracking experiments for HDR videos

This dataset is used for action localization tasks.

EPIC-KITCHENS is a large-scale egocentric video benchmark recorded by 32 participants in their native kitchen environments. Our videos depict non-scripted daily activities: we...

RF PIX2PIX Unsupervised Wi-Fi to Video Translation

The KTH action dataset consists of humans performing 6 types of actions: boxing, clapping, waving, jogging, running, and walking under 4 scenarios: outdoors, outdoors with scale...

The Chinese Sign Language (CSL) dataset is a dataset of Chinese sign language videos. It contains 100 Chinese daily expressions, each demonstrated 5 times by 50 presenters.

A benchmark dataset and evaluation methodology for video object segmentation.

Amazon-Video dataset is a dataset of user-item interactions from the Amazon video streaming service, which is used for evaluating the performance of recommendation algorithms.

Feature shifts have been shown to be useful for action recognition with CNN-based models since Temporal Shift Module (TSM) was proposed. It is based on frame-wise feature...

The VGGSound dataset is a large-scale audio-visual dataset containing 10,000 10-second video clips with corresponding audio files.

Video super-resolution (VSR) is the technique of increasing the nominal resolution of image/video content accompanied with quality improvement.