A novel and fast shape classification and regularization algorithm for on-line sketchy graphics recognition is proposed. We divide the on-line graphics recognition process into four stages: preprocessing,shape classif...
详细信息
A novel and fast shape classification and regularization algorithm for on-line sketchy graphics recognition is proposed. We divide the on-line graphics recognition process into four stages: preprocessing,shape classification,shape fitting,and regularization. Attraction Force Model is employed to progressively combine the vertices on the input sketchy stroke and reduce the total number of vertices before the type of shape can be determined. After that ,the shape is fitted and gradually rectified to a regular one,thus the regularized shape fits the user intended one *** results show that this algorithm can yield good recognition precision(averagely above 90% )and fine regularization effect but with fast speed. Consequently,it is especially suitable to computational critical environment such as PDAs,which solely depends on a pen-based user interface.
Interpreting background is very important in natural scene images. This paper addresses the automaticclassification of background region by using visual semantic template. The method to create the templates is intro-d...
详细信息
Interpreting background is very important in natural scene images. This paper addresses the automaticclassification of background region by using visual semantic template. The method to create the templates is intro-duced first. Then we use these templates to classify background regions, and the results are analyzed. We also usethese templates to locate background objects in images, and to determine whether an image contains certain kind ofbackground. The result is promising in object locating. Some approaches to improve the ability of these templates arealso discussed.
随着Internet上信息量的飞速增长,成千上万的网上文档需要分类以方便用户的浏览和获取。因此文档的自动分类工作已经越来越受到重视,一些相应的分类方法也应运而生。但其中很少有涉及到"层次化"的分类领域,且绝大多数方法仅仅返回单个分类结果。文中,我们提出了一种新的文档自动分类方法:MRHC(Multicategory ReturnedAlgorithmforHierarchicalClassification)。该方法着眼于层次化的分类技术,并在适当的情况下为文档返回多个分类结果。该方法中结合了特征削减和增量学习技术以便提高分类性能。最后,为了更加准确、客观的评价分类结果,提出了一种新的评估方法:LEP(Length of Error Path)。实验结果表明,提出的分类方法响应时间短,分类准确度高,具有较强的实用性。
暂无评论