IVML  
  about | r&d | publications | courses | people | links
   

Y. Xirouhakis, A. Tirakis and A. Delopoulos
An Efficient Graph Representation for Image Retrieval based on Color Composition
Third IMACS/IEEE Conference on Circuits, Systems, Communications and Computers (CSCC'99), Athens, Greece
ABSTRACT
Content-based video/image indexing and retrieval is of paramount importance in modern multimedia applications mainly due to the large amount of information involved. Querying by visual example, where the user provides an example image, and querying by sketch, employing user-constructed sketches, are important options in such systems. Prototype systems providing content-based image query and retrieval capabilities have been reported in literature, enabling efficient browsing in multimedia databases. In these systems, content information is modeled in terms of color, texture and shape attributes. However, existing techniques on image similarity retrieval, based on general color attributes, as well as localized color, have generally proved not to yield adequate results. Present work focuses on the extraction of image features in terms of their color composition information using a graph-based approach. This approach inherently captures local color information, and provides improved results in both the case of seeking images similar to an example image and the case of images containing a part similar to the image at hand. Experiments have been included to verify the efficiency of the algorithm. The proposed technique is evaluated in the framework of the Esprit project under development DiVAN.
01 January , 1999
Y. Xirouhakis, A. Tirakis and A. Delopoulos, "An Efficient Graph Representation for Image Retrieval based on Color Composition", Third IMACS/IEEE Conference on Circuits, Systems, Communications and Computers (CSCC'99), Athens, Greece
[ save PDF] [ BibTex] [ Print] [ Back]

© 00 The Image, Video and Multimedia Systems Laboratory - v1.12