Monotonic Search Networks For Computer Vision Databases
Donald W. Dearholt, Naomi Gonzales, Gayathri Bindu Kurup · 2005
Design criteria for an associative network database for com- puter vision which supports monotonic search for the best match of an unknown entity are discussed. The network paradigm used provides for associative clustering of similar entities, and thus higher levels of abstraction are supported. A small database using thirty entities, each represented by a Fourier feature vector, is constructed and tested.