Abstract:ABSTRACTEvent detection plays an essential role in video content analysis and remains a challenging open problem. In particular, the study on detecting human-related video events in complex scenes with both a crowd of people and dynamic motion is still limited. In this paper, we investigate detecting video events that involve elementary human actions, e.g. making cellphone call, putting an object down, and pointing to something, in complex scenes using a novel spatio-temporal descriptor based approach. A new spatio-temporal descriptor, which temporally integrates the statistics of a set of response maps of low-level features, e.g. image gradients and optical flows, in a space-time cube, is proposed to capture the characteristics of actions in terms of their appearance and motion patterns. Based on this kind of descriptors, the bag-of-words method is utilized to describe a human figure as a concise feature vector. Then, these features are employed to train SVM classifiers at multiple spatial pyramid levels to distinguish different actions. Finally, a Gaussian kernel based temporal filtering is conducted to segment the sequences of events from a video stream taking account of the temporal consistency of actions. The proposed approach is capable of tolerating spatial layout variations and local deformations of human actions due to diverse view angles and rough human figure alignment in complex scenes. Extensive experiments on the 50-hour video dataset of TRECVid 2008 event detection task demonstrate that our approach outperforms the well-known SIFT descriptor based methods and effectively detects video events in challenging real-world conditions.

A System Based On Sequence Learning For Event Detection In Surveillance Video

Esur: A System for Events Detection in Surveillance Video

Video Based Head Detection and Tracking Surveillance System

PKU-NEC @TRECVID2012 SED - Uneven-Sequence Based Event Detection in Surveillance Video.

Detecting Human Actions in Surveillance Videos

PKU@TRECVID2010: Pair-Wise Event Detection in Surveillance Video.

BIT @ TRECVid 2013 : Surveillance Event Detection

PKU@TRECVID2009: Single-Actor and Pair-Activity Event Detection in Surveillance Video

A Comprehensive Solution for Detecting Events in Complex Surveillance Videos

Surveillance Event Detection Based on CNN

A Unified Framework with a Benchmark Dataset for Surveillance Event Detection.

Interactive Surveillance Event Detection through Mid-level Discriminative Representation

Pair-wise event detection using cubic features and sequence discriminant learning

Robust Event Detection Scheme for Complex Scenes in Video Surveillance

Informedia@TRECVID 2011: Surveillance Event Detection.

TRECVID 2008 Event Detection By MCG-ICT-CAS *

Detecting video events based on action recognition in complex scenes using spatio-temporal descriptor.

A Spatio-Temporal Deep Architecture for Surveillance Event Detection Based on ConvLSTM

Early Event Detection Based on Dynamic Images of Surveillance Videos

TJUT-TJU@TRECVID 2011: Surveillance Event Detection.

A Framework for an Event Driven Video Surveillance System