A model for the qualitative description of images based on visual and spatial features

An approach that provides a qualitative description of any image is presented in this paper. The main visual features (shape and colour) and the main spatial features (fixed orientation, relative orientation and topology) of each object within the image are described. This approach has been tested in two real scenarios that involve agents and human interaction: (i) images captured by the webcam of a mobile robot while it navigates, and (ii) images of tile compositions captured by an industrial camera used to select tile pieces to be used in assembling tile mosaics. In both scenarios, promising results have been obtained.

Figure optionsDownload high-quality image (81 K)Download as PowerPoint slideHighlights
► A computational approach for qualitative description of any digital image.
► All the characteristic regions within the image are described visually and spatially.
► Qualitative models of shape, colour, topology, and fixed and relative orientation.
► Approach independent of the image segmentation method used.
► Successful results obtained with images of indoor environments and tile compositions.