Title :
Fuzzy aggregation of palette colors for hybrid querying of fine art image databases
Author :
Androutsos, P. ; Kushki, A. ; Plataniotis, K.N. ; Venetsanopoulos, A.N.
Author_Institution :
Dept. of Electr. & Comput. Eng., Toronto Univ., Ont., Canada
Abstract :
As outlined by the ISO committee, the problem of hybrid query generation lies outside the scope of the MPEG-7 standard. This problem of creating intelligent image database queries that both correctly reflect the intentions of the user as well as provide good retrieval results can be approached in various ways. This paper proposes a hybrid query generation scheme which employs fuzzy aggregation for including and excluding palette colors within the context of a fine art database containing various paintings and drawings. The aggregator herein exhibits flexibility in its logical behaviour through parameters that can be set by the designer as well permitting the exclusion of specific colors from queries. This translates to richer controls for a user wishing to locate works from a large art image database that have similar, yet complex color palettes. Experimentation on an image database of 464 paintings and drawings illustrate this fact, and a comparison with a weighted mean approach is provided.
Keywords :
art; content-based retrieval; feature extraction; image colour analysis; image recognition; image retrieval; visual databases; drawings; fine art database; fuzzy aggregation; hybrid query generation; intelligent image database queries; paintings; palette colors; Art; Color; Deductive databases; Hybrid power systems; ISO standards; Image databases; Image retrieval; Information retrieval; MPEG 7 Standard; Painting;
Conference_Titel :
Digital Signal Processing, 2002. DSP 2002. 2002 14th International Conference on
Print_ISBN :
0-7803-7503-3
DOI :
10.1109/ICDSP.2002.1027843