DocumentCode
381432
Title
Automatic distinction of camera and object motions in video sequences
Author
Oh, JungHwan ; Sankuratri, Praveen
Author_Institution
Dept. of Comput. Sci. Eng., Texas Univ., Arlington, TX, USA
Volume
1
fYear
2002
fDate
2002
Firstpage
81
Abstract
The motions in a video sequence are caused by two different sources which are camera and object(s). In other words, these motions can be divided into camera motions, object motions and combination of both. To express different motions accurately, therefore, we need to distinguish the different motions generated by the different sources. The goal of this research is to distinguish camera and object(s) motions for video content analyzing and indexing purposes. To do this, we first measure overall motions and detect camera motion changes in a shot using the techniques in our previous works, then distinguish object and camera motions from overall motions using the technique proposed in this paper. Not only the degrees (amounts) but also the locations of motions are computed and presented accurately by a two dimensional matrix. Our preliminary experimental results show the possibility that the proposed technique can distinguish these motions as expected.
Keywords
image sequences; object detection; video signal processing; automatic distinction; camera motions; change detection; object motions; two dimensional matrix; video sequences; Cameras; Image motion analysis; Indexing; Motion analysis; Motion detection; Motion measurement; Object detection; Optical computing; Optical filters; Video sequences;
fLanguage
English
Publisher
ieee
Conference_Titel
Multimedia and Expo, 2002. ICME '02. Proceedings. 2002 IEEE International Conference on
Print_ISBN
0-7803-7304-9
Type
conf
DOI
10.1109/ICME.2002.1035723
Filename
1035723
Link To Document