| 摘 要: 在图像检索中，如何有效提取图像特征是基于内容的图像检索中的一个难点。针对该难题，提出了一种 基于遗传算法的图像多特征权重自动赋值方法。首先使用灰色直方图提取颜色特征并利用树形小波提取纹理特征，然后 利用遗传算法的全局最优解搜索功能自动确定各特征的权重。实验结果分析表明：在灰度图像的相似性检索中，基于遗 传算法的多特征权重自动赋值方法与其他方法相比，平均查全率增加将近8%，平均查准率增加将近9%，说明该方法有 较高的检索精确度。 |
| 关键词: 树型小波；特征融合；遗传算法；图像检索 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 湖南省自然科学基金(2018JJ4068)；湖南省高铁运行安全保障工程技术研究中心开放基金(2017TP2022-17KJ104). |
|
| An Image Multi-Feature Weight Automatic Assignment Method Based on Genetic Algorithm |
|
ZHANG Xiaoli,XIAO Mansheng,YE Zixuan
|
( School of Computer Science, Hunan University of Technology, Zhuzhou 412007, China)
|
| Abstract: In image retrieval,how to effectively extract image features is a difficult point in content-based image retrieval.Aiming at this problem,the paper proposes an automatic multi-feature weight assignment method based on genetic algorithm.First,the method uses gray histogram to extract color features and uses tree wavelet to extract texture features.Then the genetic algorithm's global optimal solution search function is used to automatically determine the weight of each feature. The experimental results show that in the similarity search of gray image,the multi-feature weight automatic assignment method based on genetic algorithm has increased the recall rate by 8% and increased the precision rate by 9%,compared with other methods,which proves the high retrieval accuracy of this method. |
| Keywords: tree wavelet;feature fusion;genetic algorithm;image retrieval |
