基于全局特征信息的快速场景识别与分类研究

Rapid Scene Recognition and Classification Based on Global Characteristics

全局特征信息又称为"Gist"信息,为场景的低维签名向量。采用全局特征信息对场景进行识别与分类不需要对图像进行分割和局部特征提取,可以实现快速场景识别与分类。本论文阐述了基于全局特征的场景识别与分类的一些重要研究成果,分析比较了各个"Gist"模型的特征,并对多个场景进行快速识别和分类,分析比较了实验研究成果,最后对快速场景识别与分类的研究前景做了展望。

＂Gist＂ is the global characteristics information of a scene,it＇s a low-dimensional signature vector of a scene.The key researches of scene recognition and classification based on ＂Gist＂ are reviewed in this paper.Some experiments are carried out based on the ＂Gist＂ models,and they are applied to classify three classes of scenes.The prospect of rapid scene recognition and classification is also introduced in conclusion.