79 datasets found

CityCam dataset contains recordings from 212 traffic cameras in the United States, with a resolution of 352x240 and a frame rate of 1 frame/per second.

Visual object tracking, which tracks a specified target in a changing video sequence automatically, is a fundamental problem in many topics such as visual analysis, automatic...

A fine-grained crowd behavior analysis dataset with 16 videos and annotated crowd behaviors.

A crowd emotion dataset with 31 videos and annotated crowd emotions.

A crowd counting dataset with over 1000 labeled videos captured by over 100 monitoring cameras.

A crowd video captioning dataset based on the WorldExpo'10 dataset, with 98 videos selected and captions generated for them.

A large-scale database for partial copy detection in videos.

The dataset is used for breast cancer diagnosis in ultrasound videos.

Tinyvideos dataset

The STV-IDL dataset is a collection of videos with referring expressions that describe spatio-temporal events.

The VisDrone2019 dataset contains 288 video clips made up of 261,908 frames and 10,209 images

A dataset of 30 videos of length 3 minutes to 20 minutes from five classes of daily activities: socializing, home repair, biking around urban areas, cooking, and home tours.

The ActivityNet1.2 dataset is a large-scale benchmark for action recognition and localization in videos.

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...

Temporal sentence grounding in videos (TSGV) is a task to retrieve a video segment that semantically corresponds to a query in natural language.

A challenging real-world dataset for deepfake detection.

A deepfake detection method that can address the issue of modifying segments of videos using generative techniques.

The AGRE ADOS database, Kaggle database, and a self-gathered video test dataset with corresponding ADOS data

The HMDB dataset consists of 6766 videos from 51 different action categories. The videos are generally of low quality, with strong camera motion, and non-centered people.

InternVid: A large-scale video-text dataset for multimodal understanding and generation.