TY - GEN
T1 - A context-based whole video retrieval system with dynamic video stream analysis mechanisms
AU - Kurabayashi, Shuichi
AU - Ueno, Taichi
AU - Kiyoki, Yasushi
PY - 2009
Y1 - 2009
N2 - With the rapid expansion of fictional video contents on WWW, it is important to develop video retrieval mechanisms for fictional video contents, such as TV drama, animations, and movies. This paper proposes a context-based video retrieval system for retrieving an entire video stream, not individual scenes or fragments. The unique feature of this system is a SQL-like contextual multimedia query language, which allows users to express and characterize their queries by using video clips as contexts. It extracts emotive contexts of visual elements from video files by analyzing the coloremotions hidden within video content using methods developed during color psychology research. The system's capabilities enables new ways of querying artistic, visually rich and dynamic video content - something which is not possible by using just a few words. This is done by inspecting the dynamic and time-evolving contexts of video data. A prototype system has been implemented which is applicable to performing queries against television content. This paper clarifies the feasibility of the system by demonstrating the results of several experiments.
AB - With the rapid expansion of fictional video contents on WWW, it is important to develop video retrieval mechanisms for fictional video contents, such as TV drama, animations, and movies. This paper proposes a context-based video retrieval system for retrieving an entire video stream, not individual scenes or fragments. The unique feature of this system is a SQL-like contextual multimedia query language, which allows users to express and characterize their queries by using video clips as contexts. It extracts emotive contexts of visual elements from video files by analyzing the coloremotions hidden within video content using methods developed during color psychology research. The system's capabilities enables new ways of querying artistic, visually rich and dynamic video content - something which is not possible by using just a few words. This is done by inspecting the dynamic and time-evolving contexts of video data. A prototype system has been implemented which is applicable to performing queries against television content. This paper clarifies the feasibility of the system by demonstrating the results of several experiments.
KW - Context-computing
KW - Multimedia database
KW - Query language
KW - Video search engine
KW - Video stream analysis
UR - http://www.scopus.com/inward/record.url?scp=77949539552&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77949539552&partnerID=8YFLogxK
U2 - 10.1109/ISM.2009.85
DO - 10.1109/ISM.2009.85
M3 - Conference contribution
AN - SCOPUS:77949539552
SN - 9780769538907
T3 - ISM 2009 - 11th IEEE International Symposium on Multimedia
SP - 505
EP - 510
BT - ISM 2009 - 11th IEEE International Symposium on Multimedia
T2 - 11th IEEE International Symposium on Multimedia, ISM 2009
Y2 - 14 December 2009 through 16 December 2009
ER -