10 datasets found

Creating summaries from user videos. The SumMe dataset consists of 25 videos (with the average length of 160 seconds) and 50 videos (with the average length of 252 seconds)...

The dataset is a collection of YouTube trending videos with their thumbnails, metadata, and object labels.

The ChaLearn First Impressions v2 dataset is a collection of 10000 HD 720p YouTube videoclips, gathered from 3060 unique videos.

The How2 dataset consists of summarizations of How2 videos taken from YouTube.

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

The dataset used in the paper is a large-scale graph dataset, consisting of users and shows with multi-attribute edges. The graph is constructed by selecting user IDs and side...

Automatic keyframe detection from videos is an exercise in selecting scenes that can best summarize the content for long videos.

The YouTube-VOS 2018 dataset is a large-scale benchmark for video object segmentation.

Motion has shown to be useful for video understanding, where motion is typically represented by optical flow. However, computing flow from video frames is very time-consuming....

Speaker verification systems experience significant performance degradation when tasked with short-duration trial recordings. To address this challenge, a multi-scale feature...