节点文献
基于Sift特征的服装纸样检索
Pattern retrieval algorithm based on Sift feature
【摘要】 提出了基于Sift特征的服装纸样检索方法.首先提取服装结构图的Sift特征,再采用近似的最近邻搜索算法在纸样数据库中检索相似纸样.该方法对于相似的图像或子图像具有良好的识别能力,可大大减少服装纸样检索的时间,从而有效利用已有的纸样库资源,避免重复打板.另外,该方法还可用于甄别服装板型结构设计的原作.
【Abstract】 The way to search for the same draft based on Sift feature is proposed.By use of DoG,Sift consists of adjacent scale Gaussian differential unclear and input image convolution,the calculation of scale space is simplified.It has a good ability to identify the similar image or sub-image,it will save a lot time,so that we can use the source of draft best,avoiding doing the same draft and wasting time.In addition,the method is significantly used to identify clothing original design of plate-type pattern.
【关键词】 Sift算法;
Sift特征匹配;
纸样检索;
服装纸样;
【Key words】 Sift; Sift feature matching; pattern retrieval; garment pattern;
【Key words】 Sift; Sift feature matching; pattern retrieval; garment pattern;
- 【文献出处】 天津工业大学学报 ,Journal of Tianjin Polytechnic University , 编辑部邮箱 ,2010年03期
- 【分类号】TS941.2
- 【被引频次】6
- 【下载频次】109