基于多尺度上下文语义信息的图像场景分类算法

摘要

传统视觉词典模型没有考虑图像的多尺度和上下文语义共生关系。本文提出一种基于多尺度上下文语义信息的图像场景分类算法。首先，对图像进行多尺度分解，从多个尺度提取不同粒度的视觉信息；其次利用基于密度的自适应选择算法确定最优概率潜在语义分析模型主题数；然后，结合 Markov 随机场共同挖掘图像块的上下文语义共生信息，得到图像的多尺度直方图表示；最后结合支持向量机实现场景分类。实验结果表明，本文算法能有效利用图像的多尺度和上下文语义信息，提高视觉单词的语义准确性，从而改善场景分类性能。%The conventional BoVW neglects image multi-scale and contextual semantic co-occurrence information .This pa-per proposes an image scene classification algorithm based on multi-scale and contextual semantic information .Firstly ,Images are decomposed into variant scales and diverse visual details are extracted from different scale layers .Secondly ,a density-based adaptive selection method is employed to choose the best topic number of probabilistic latent semantic analysis model .Then ,Markov random field are combined to mine the contextual semantic co-occurrence information ,thus to obtain a multi-scale histogram as the image representation .Finally ,the support vector machine classifier is utilized to perform scene classification .The experimental results demonstrate that our algorithm can effectively utilize the multi-scale and contextual semantic information of images and improve im-age scene classification performance .