基于分块直方图和共生矩阵的图像检索方法

Image retrieval based on blocked histograms and co-occurrence matrix

提出了一种基于分块直方图和共生矩阵相结合的图像检索方法。传统的基于颜色直方图特征的图像检索方法简单且易于实现,但是失去了颜色的空间分布信息。分块直方图法根据图像目标和背景的不同,将图像分成意义不同的8个子块,充分利用了图像分块的特点,改善了颜色特征缺乏空间信息的缺点。共生矩阵法根据不同图像间共生矩阵的差别来实现图像检索。最后,综合利用上述两个特征对图像进行共同检索。实验表明:综合2种特征检索图像比单一特征检索效果更好。

A new method for image retrieval which was based on both blocked histograms and co-occurrence matrix was proposed. It was easy to perform for traditional image retrieval based on color histogram, however, its space distribution information was lost. Blocked histograms method was a feature extraction method based on the differences between the object and background, which divided the image into 8 different sub-blocks. This method amended the disadvantage of lacking spatial information of common color feature. The co-occurrence matrix was different for different image texture,and It can be used for performing the image retrieval based on co-occurrence matrix. At last, a new method for image retrieval using combined two features was proposed. The result showed that the retrieval results obtained from combined--features are superior to that obtained from single-features.