V3C1 Dataset

Abstract

In this work we analyze content statistics of the V3C1 dataset, which is the first partition of the Vimeo Creative Commons Collection (V3C). The dataset has been designed to represent true web videos in the wild, with good visual quality and diverse content characteristics, and will serve as evaluation basis for the Video Browser Showdown 2019-2021 and TREC Video Retrieval (TRECVID) Ad-Hoc Video Search tasks 2019-2021. The dataset comes with a shot segmentation (around 1 million shots) for which we analyze content specifics and statistics. Our research shows that the content of V3C1 is very diverse, has no predominant characteristics and provides a low selfsimilarity. Thus it is very well suited for video retrieval evaluations as well as for participants of TRECVID AVS or the VBS.

Cottontail DB: An Open Source Database System for Multimedia Retrieval and Analysis

“…To demonstrate Cottontail DB's ability to cope with large multimedia collections, we compare ADAM pro and Cottontail DB in terms of retrieval speed for nearest neighbour queries. We base the evaluation on the V3C1 [1] video collection, which has an overall collection size of = 1.088.896 segments and is used in the VBS setting and therefore representative of interactive video retrieval tasks. For the comparison, we generated features using Cineast [15] with vector dimensions ranging from = 3 to = 2048 and compared them using the Euclidean (L2) distance.…”

A VR Interface for Browsing Visual Spaces at VBS2021

“…In this task, we use the first part of the Vimeo Creative Commons Collection dataset 1 (V3C) called V3C1 [1], a wide-ranging video collection. V3C1 dataset includes 7,475 videos (1,000 hours), categorized into different content categories ranging from Food, Fashion, Art to Instructional videos.…”

VISIONE at Video Browser Showdown 2023

“…VBS is an international video search competition that is held annually since 2012 [13]. The V3C1 dataset [5], consisting of 7,475 videos, has been used in the competition since 2019. So far, three types of search tasks are considered in the competition: Known-Item-Search (KIS), textual KIS and Ad-hoc Video Search (AVS).…”