【摘要】 图像数据库的不断庞大使基于内容的图像检索成为研究热点,目前主要集中于底层特征的相似度匹配。文章重点介绍了颜色特征中的分块主色法,纹理特征中的灰度共生矩阵法和形状特征中的小波变换及不变矩法。在利用单一特征检索的基础上,该文提出了一种综合利用上述三种特征共同进行检索的方法。同时,还将相关反馈技术融合到算法中,通过权值矩阵的正负调整及三种特征系数的调整来提高检索准确率。由实验数据表明,文中的方法是很有效的。更多还原

【Abstract】 Content-based image retrieval becomes a hot research topic due to the rapid increase of image database,and the present explore focuses on the similarity matching of lower-features.In this paper,the color features are represented with the dominant color of partition and the texture features are calculated with the co-occurrence matrices.Wavelet transform is used in image edge detection and affine immutability of invariant moment is used in the feature extraction of the shape.On the basis of using color and texture features separately,a new method for image retrieval using color and texture and shape features is proposed.Meanwhile,the technique of relevance feedback is used in the algorithm to enhance the effectiveness of retrieval by adjusting the value matrix and feature parameters.Experiment results show that the proposed approach is effective.更多还原