基于模糊区域特征的图像检索算法和关联反馈算法是当前图像检索领域的研究热点,由于区域模糊相似度的复杂性,绝大多数关联反馈算法不能应用到基于模糊区域特征的图像检索方法中。为解决这个问题,论文修改了模糊相似度计算方法,并结合经典的基于权重调整的关联反馈算法,提出一种基于模糊区域特征的关联反馈算法。对脑出血CT图像的检索实验结果表明该算法效果较好。
Fuzzy region based image retrieval algorithms and relevance feedback algorithms have been active research areas. However,most relevance feedback algorithms can't be used in fuzzy region based image retrieval system because of the complexity of fuzzy similarity expression.To overcome the problem,this paper modifies the equation to calculate the fuzzy similarity and proposes a new relevance feedback method based on fuzzy features and the classical re-weighting relevance feedback algorithm. Cerebral hemorrhage CT image retrieval experiments show that the algorithm is efficient.