Detection of Abrupt Changes in Spatial Relationships in Video Sequences

Abstract :

Detecting unusual events in video sequences is very challenging due to cluttered background, the difficulties of accurate extraction and tracking of moving objects, illumination change, etc. In this work, we focus on detecting strong changes in spatial relationships between moving objects in video sequences, with a limited knowledge of the objects. In this approach, the spatial relationships between two objects of interest are modeled using angle and distance histograms as examples. To evaluate the evolution of the spatial relationships during time, the distances between two angle or distance histograms at two different instants in time are estimated. In addition, a combination approach is proposed to combine the evolution of directional (angle) and metric (distance) relationships. Studying the evolution of the spatial relationships during time allows us to detect the ruptures in such spatial relationships. This study can constitute a promising step toward event detection in video sequences, with few a priori models on the objects.

Document type :
Book sections
Complete list of metadatas

https://hal.telecom-paristech.fr/hal-02287241
Contributor : Telecomparis Hal <>
Submitted on : Friday, September 13, 2019 - 4:45:41 PM
Last modification on : Thursday, October 17, 2019 - 12:37:00 PM

Identifiers

  • HAL Id : hal-02287241, version 1

Collections

Citation

Abdel-Bassir Abou-Elailah, Valérie Gouet-Brunet, Isabelle Bloch. Detection of Abrupt Changes in Spatial Relationships in Video Sequences. Pattern Recognition, Applications and Methods, LNCS 9493, Springer, pp.89-106, 2015. ⟨hal-02287241⟩

Share

Metrics

Record views

7