Similarity engine for content-based retrieval of images

作者: Bradley Horowitz , Jeffrey R. Bach , Chiao-fe Shu , Ramesh Jain , Amarnath Gupta

DOI:

关键词: Modular designSchema (psychology)Feature vectorComputer visionContent based retrievalComputer scienceVisual ObjectsArtificial intelligence

摘要: A system and method for content-based search retrieval of visual objects. base information (VIR) engine utilizes a set universal primitives to operate on the An extensible VIR allows custom, modular be defined registered. custom primitive addresses domain specific problems can utilize any image understanding technique. Object attributes extracted over entire or only portion object. schema is as collection primitives. implies features processed corresponding feature vector used similarity scoring. registration interface registers facilitates storing an analysis function comparison table. heterogeneous objects analyzed by different schemas compared if at least one in common between schemas. threshold-based utilized improve performance engine. distance two vectors computed processes so generate score.

参考文章(7)
Hiroyuki Shimogaki, Takio Kurita, Toshikazu Kato, Tetsuya Mizutori, Koreaki Fujimura, A cognitive approach to visual interaction International conference on Multimedia information systems '91. pp. 109- 122 ,(1991)
William Robinson Equitz, Peter Cornelius Yanker, Myron Dale Flickner, Carlton Wayne Niblack, Dragutin Petkovic, Ronald Jason Barber, Bradley James Beitel, Thomas Randolph Work, Image query system and method ,(1994)
Kwang-Soo Hahn, A. Kathleen Hennessey, YouLing Lin, Apparatus and method for image processing in symbolic space ,(1994)
Toshikazu Kato, Database architecture for content-based image retrieval SPIE/IS&T 1992 Symposium on Electronic Imaging: Science and Technology. ,vol. 1662, pp. 112- 123 ,(1992) , 10.1117/12.58497