A Large-scale Benchmark Dataset for Event Recognition in Surveillance Video

DMCA

A Large-scale Benchmark Dataset for Event Recognition in Surveillance Video

Cached

Download Links

by Sangmin Oh , Anthony Hoogs , Amitha Perera , Naresh Cuntoor , Chia-chih Chen , Jong Taek Lee , Saurajit Mukherjee , J. K. Aggarwal , Hyungtae Lee , Larry Davis , Eran Swears , Xioyang Wang , Qiang Ji , Kishore Reddy , Mubarak Shah , Carl Vondrick , Hamed Pirsiavash

Citations:

44 - 9 self

BibTeX

@MISC{Oh_alarge-scale,
    author = {Sangmin Oh and Anthony Hoogs and Amitha Perera and Naresh Cuntoor and Chia-chih Chen and Jong Taek Lee and Saurajit Mukherjee and J. K. Aggarwal and Hyungtae Lee and Larry Davis and Eran Swears and Xioyang Wang and Qiang Ji and Kishore Reddy and Mubarak Shah and Carl Vondrick and Hamed Pirsiavash},
    title = {A Large-scale Benchmark Dataset for Event Recognition in Surveillance Video},
    year = {}
}

OpenURL

Abstract

We introduce a new large-scale video dataset designed to assess the performance of diverse visual event recognition algorithms with a focus on continuous visual event recognition (CVER) in outdoor areas with wide coverage. Previous datasets for action recognition are unrealistic for real-world surveillance because they consist of short clips showing one action by one individual [15, 8]. Datasets have been developed for movies [11] and sports [12], but, these actions and scene conditions do not apply effectively to surveillance videos. Our dataset consists of many outdoor scenes with actions occurring naturally by non-actors in continuously captured videos of the real world. The dataset includes large numbers of instances for 23 event types distributed throughout 29 hours of video. This data is accompanied by detailed annotations which include both moving object tracks and event examples, which will provide solid basis for large-scale evaluation. Additionally, we propose different types of evaluation modes for visual recognition tasks and evaluation metrics along with our preliminary experimental results. We believe that this dataset will stimulate diverse aspects of computer vision research and help us to advance the CVER tasks in the years ahead. 1.

Keyphrases

event recognition large-scale benchmark dataset surveillance video wide coverage previous datasets preliminary experimental result large-scale evaluation action recognition continuous visual event recognition evaluation mode real-world surveillance diverse aspect different type detailed annotation new large-scale video dataset scene condition computer vision research many outdoor scene cver task short clip diverse visual event recognition algorithm event example large number object track event type solid basis evaluation metric visual recognition task real world outdoor area

DMCA