摘要
提出了一种新的基于形状的图像检索方法 .首先利用广义模糊算子对图像进行边缘检测 ,以检测出的边缘映射图计算不变矩来代表边缘图像的形状信息 ,从而完成对图像的自动索引 ;其次 ,重新定义了相似度测度 ,提高了匹配速度 ;最后 ,设定纵横比辅助搜索 ,以提高检索精度 .实验结果表明 ,此方法在一定程度上排除了与形状无关信息的影响 ,不但简单有效 ,而且搜索速度快 ,具有一定的实用意义 .
A new method for shape_based image retrieval is presented. Firstly the edges in the original image using generalized fuzzy operator (GFO) are detected, the moment invariants of the edge map are calculated, and the moment invariants as the shape features are automatically indexed. Then, the similarity metrics to acelerate the match are redefined. Finally, the aspect ratio was defined to improve the retrieval precision. Experimental results show that the new method can remove some irrelevant information related to shape in the original image and weaken its influnce on the retrieval results to some extent. This new method is simple, effctive and efficient and it is of some practical significance.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2002年第4期60-64,共5页
Journal of South China University of Technology(Natural Science Edition)
基金
高等学校骨干教师资助计划项目 (教技司[2 0 0 0 ]6 5号 )
高等学校优秀青年教师教学科研奖励计划资助项目
关键词
GFO
不变矩
形状
图像检索
image retrieval
shape
moment invariants
generalized fuzzy operator
