An experimental comparison of machine learning for adaptive sketch recognition

摘要

Benefiting from advances in online sketchy shape recognition system could not be expected before the problem of its adaptability is well solved. In this paper, three methods of machine learning are compared for adaptive sketch recognition with some experiments based on various feature representations of sketches and collected samples of multi-users. The experimental results reveal elementally some important matters of sketch recognition based on machine learning methods and have a significant reference to further researches of adaptive sketch recognition. (c) 2006 Elsevier Inc. All rights reserved.