Techniques for managing multi-dimensional spatial data are disclosed. The multi-dimensional spatial data include, but are not limited to, two- and three-dimensional imagery data, video clips, motion picture files and movies. In particular, the present invention enables the user to match sub-regions of query data with stored spatial data in a database using foveal indices. According to one aspect of the present invention, the multi-dimensional spatial data is encoded with the following technique, which segments the spatial data into overlapping samples (“foveal samples”), represents foveal samples as polar-domain foveals in a polar coordinate system, processes the foveals into a scale- and rotation-invariant canonical form (“canonical foveal”), extracts features from foveals using a mathematical transform (e.g., wavelet, beamlet), and generates and stores a set of foveal indices for each canonical foveal matching application criteria. It is emphasized that the present invention applies to multi-dimensional spatial data.