Logo recognition based on a novel pairwise classification approach

摘要: Logo recognition is an important task in the field of document image processing and retrieval. Successful logos facilitates automatic classification source documents, which has been considered as a key strategy for analysis. From machine learning point view, logo may be multi-class problem. In this paper, novel pairwise method proposed applied to application. The system takes advantages simplicity speed nearest neighbor algorithm strength other powerful binary classifiers discriminate between two classes. first validated on set UCI Machine Learning Repository datasets then real vision experimental results show that technique not only achieves better accuracy, but also computationally more efficient tackling problems have large number target