SVM Based Shot Boundary Detection Using Block Motion Feature Based on Statistical Moments

Author

Bhowmick, Brojeshwar ; Goswami, Kaustav

Author_Institution

Innovation Lab., Tata Consultancy Services Ltd., Kolkata

fYear

2009

fDate

4-6 Feb. 2009

Firstpage

134

Lastpage

137

Abstract

Temporal video segmentation is of fundamental importance in order to facilitate userpsilas access to huge volume of video data as well as for video summarization.The objective of shot boundary detection is to partition the video into meaningful, basic structural units called shots. In this paper, a shot boundary detection technique has been proposed for cuts. The method extracts block feature based similarities from the frames of the input video. Statistical moments up to second order are used to measure the motion present in the frames. Feature vectors are generated using a sliding window over time and are trained by a SVM to identify the cuts.

Keywords

feature extraction; image motion analysis; image segmentation; learning (artificial intelligence); statistical analysis; support vector machines; video signal processing; SVM training; block motion feature extraction; shot boundary detection; sliding window; statistical moment; support vector machine; temporal video segmentation; video summarization; Cameras; Feature extraction; Gunshot detection systems; Histograms; Kernel; Layout; Motion detection; Support vector machines; Technological innovation; Video sequences; cut; shot; temporal video segmentation;

fLanguage

English

Publisher

ieee

Conference_Titel

Advances in Pattern Recognition, 2009. ICAPR '09. Seventh International Conference on

Conference_Location

Kolkata

Print_ISBN

978-1-4244-3335-3

Type

conf

DOI

10.1109/ICAPR.2009.25

Filename

4782759