一种类似视频场景分割的家庭数字照片分类方法研究

Research on a New Home Digital Photo Classification Algorithm Similar to Video Scene Segmentation

在基于内容的图像分类研究中借鉴了视频场景分割的思想，提出了一种针对家庭数字照片的基于内容的全新的分类方法。该方法能自动提取种子照片并局部调整，然后进行照片聚类，实现了照片自动分类。本算法已经用Visual C++在Windows 2000环境下实现，实验结果表明，实际效果良好,且具有较强的鲁棒性。

This paper presents a novel image classification method similar to the video scene segmentation. During the image classification process, the seed images are selected from image set automatically. The method is implemented under Windows 2000 using Visual C++ . The method is robust and quite satisfying experiment results have been attained.