基于SVM的相关反馈图像分类和检索方案

摘要

图像获取和存储技术的进步可以获得包含大量有用信息的图像数据,在传统的图像分类和检索方案中,图像的低级视觉特征和高级概念之间存在着较大的语义间隔,导致图像的分类和检索效果不佳.针对该问题,提出了一种基于SVM相关反馈的图像分类和检索方案.该方案通过缩窄图像的领域,利用机器学习方法建立图像类的模型,并使用一种优化的SVM相关反馈图像检索方法学习图像的类别,将学习到的模型用于图像的分类和检索.实验结果表明,此方案能够高效的检索出较多相关图像,并对其进行有效分类.%A lot of image dates with many useful information could be got because of the technologies of advanced image acquisition and storage. In the traditional approach of image classification and retrieval, there was a wide semantic gap between the low level features and the high level concepts. To solve the problem, the image classification and retrieval scheme based on SVM was proposed. This scheme was to narrow the image domain, use machine learning methods to construct model for image classes, learn the image classification by using retrieval method of relative and feedback based on SVM, use the model to classify and retrieve the image. The experimental results showed that this scheme could be used to retrieve more relevant image, and assort it effectively.