Image clustering using content-based techniques

Abstract

The content-based QBIC information retrieval system is investigated regarding its application in rock core and textile images clustering. It is demonstrated that content-based techniques, and texture in particular, can be used to cluster images, giving a reasonable correlation with assignments made by visual inspection. Options for improving clustering accuracy include: synthesis of purer examples; outlining regions within the image; appropriate weighting of the components of the texture feature vector; and experimenting with alternative classifiers, particularly where texture is important.