Incremental Learning of Perceptual Categories for Open-Domain Sketch Recognition

Andrew Lovett, Morteza Dehghani, Kenneth Forbus

Most existing sketch understanding systems require a closed domain to achieve recognition. This paper describes an incremental learning technique for open-domain recognition. Our system builds generalizations for categories of objects based upon previous sketches of those objects and uses those generalizations to classify new sketches. We represent sketches qualitatively because we believe qualitative information provides a level of description that abstracts away details that distract from classification, such as exact dimensions. Bayesian reasoning is used in building representations to deal with the inherent uncertainty in perception. Qualitative representations are compared using SME, a computational model of analogy and similarity that is supported by psychological evidence, including studies of perceptual similarity. We use SEQL to produce generalizations based on the common structure found by SME in different sketches of the same object. We report on the results of testing the system on a corpus of sketches of everyday objects, drawn by ten different people.