融合图像显著区域二维形状特征的图像检索

Image retrieval based on two-dimensional shape features of salient region

为研究在基于内容的图像检索中形状特征对图像检索效果的影响,提出了一种融合图像显著区域二维形状特征的图像检索方法。提取图像中显著区域的形状特征,融合图像的纹理和颜色特征来对图像进行检索。当目标存在变形和大小存在巨大差异时,目前缺乏完善的数学模型来提取图像的形状特征,该方法很好地弥补了这种缺陷。通过实验结果表明,该方法能够提高图像检索的准确率。

In order to study the effect of shape features on content-based image retrieval,an image retrieval method based on two-dimensional shape features of salient region is proposed. First,the shape features of salient regions in the image are extracted; then the texture and color features of the image are merged to retrieval the image. Now,it lacks of perfect mathematical model to extract the shape of the image features when there is great difference between the deformation and size of the target,while the proposed method makes up for this defect very well. The experimental results show that the proposed method can improve the performance of image retrieval.