基于内容的索引图片自动生成算法研究

基于内容的索引图片根据原图片的感兴趣区域对原图进行索引,有效地解决了现有索引图片由于直接缩放而造成的模糊不可辨别以及失真变形等问题。采用基于内容的索引图片能大大提高检索和浏览图片信息的质量和效率。本文研究的基于内容的索引图片自动生成算法,主要是以人脸区域作为感兴趣区域,首先采用基于肤色信息和几何信息的人脸检测算法,定位出感兴趣区域;然后按比例裁剪出已检测出的人脸区域,并使用插值缩放生成基于内容的索引图片。

Search of automatic generation Algorithm of index picture based on content

The picture based on index content,which can index for the original picture according to the interested region,resolve the problem that index picture dim and distort which cause by index picture directly compressed or amplified.Adopting the index picture based on content can greatly enhance the quality and efficiency of index and view picture information.In the paper we research about the automatic generation algorithm based on content,in which we take person face as the interested region.In the algorithm,firstly we locate the interested region according to the person face detection algorithm which based on color of skin information and geometry information,and then clip the detected region of the person face to scale,and at last generate the index picture based on content by adopting the interpolation compression and amplification method.