Presentation 1994/1/21
Scene Clustering and Cut Detection in Moving Images by DCT components
Eiichi Iwanari, Yasuo Ariki,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We have studied a method to detect scene cuts in moving pictures by using DCT components.The DCT components of consecutive frames in moving pictures are similar when they are included in the same scene.Based on this fact,we propose a method to form clushters of consecutive frames by the similarity,and then to detect the scene cut when a cluster is newly formed.This method is different from detecting the change and regarding it as scene cut.By this method, detecting the scene cut becomes not so sensitive to the intensity changes which have been problems in conventional method.In this paper,we describe experimental results to find the best parameters of block size,frame width,and range of DCT components to make a clushter.Furthermore,we showed that this method is better in the case of flushing,dark moving picture and changing contrast, comparing with the conventional method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) DCT(Discrete Consine Transform) / Scene Analysis / Block devision / Cut detection
Paper # PRU93-119
Date of Issue

Conference Information
Committee PRU
Conference Date 1994/1/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Understanding (PRU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Scene Clustering and Cut Detection in Moving Images by DCT components
Sub Title (in English)
Keyword(1) DCT(Discrete Consine Transform)
Keyword(2) Scene Analysis
Keyword(3) Block devision
Keyword(4) Cut detection
1st Author's Name Eiichi Iwanari
1st Author's Affiliation Faculty of Science and Technology,Ryukoku University()
2nd Author's Name Yasuo Ariki
2nd Author's Affiliation Faculty of Science and Technology,Ryukoku University
Date 1994/1/21
Paper # PRU93-119
Volume (vol) vol.93
Number (no) 431
Page pp.pp.-
#Pages 8
Date of Issue