基于综合特征图像检索技术研究

图像数据库应用日益广泛，如何高效、准确地进行图像的检索成为一项重要的研究领域。传统的图像检索主要依赖颜色、纹理、形状、空间关系等单一视觉特征，检索效果往往不够理想。针对这一实际问题，提出一种新的图像检索方法，该方法以综合特征为检索基础，通过获得图像的归一化综合特征，并将其作为图像相似性的衡量依据。实验证明，利用综合特征进行图像检索具有很好的效果。

IMAGE RETRIEVAL TECHNIQUE BASED ON COMBINED FEATURES

The Application fields of the image database is becoming more and more widely.How to retrieval image with effective result becomes a significant research field.Traditional image retrieval methods is mainly dependent on the single vision features such as color Grain shape and so on,so its retrieval result always not ideally.To cure this problem,propose a new kind of image retrieval method which based on the feature's combination.Normalize the combined feature and use it as the standard of image's similarity metric.The test demonstrates that image retrieval based on combined feature has a very good effect.