摘要
提出了一种自动识别衣片的智能算法。衣片识别包括三个步骤:冗余曲线去除、连通区域检测以及针对每个连通区域的最大衣片识别。最大衣片识别基于最大夹角的准则,无需用户交互,并且不需遍历所有可能的封闭区域,避免了封闭区域面积的计算,提高了衣片生成的效率。实验结果表明,该算法能够实现任意复杂拓扑的衣片生成,具有很好的通用性。
An intelligent algorithm for garment identification was proposed. Garment identification consists of three steps, including removing redundant curves, detecting connected regions and identifying the largest garment for each connected region. Identifying the largest garment is based on the rule of maximum angle, that is, repeatedly find the curve which has the maximum angle with the current curve until the end vertex is equal to the start vertex. This method does not require the area computation of the closed region and traversing of all the possible closed regions. So the efficiency of garment generation is improved by a lot. Experiments show that this method can successfully generate garment with arbitrary complex topology and can be used generally in cloth design.
出处
《计算机应用》
CSCD
北大核心
2013年第A01期218-219,232,共3页
journal of Computer Applications
基金
中央高校基本科研业务费资助项目(531107040275)
湖南大学青年教师成长计划项目
关键词
衣片识别
最大夹角
连通区域
复杂拓扑
garment identification
maximum angle
connected region
complex topology
