基于PLSA-BOW模型的医学影像分类算法的研究

MEDICAL IMAGE CLASSIFICATION ALGORITHM BASED ON PLSA-BOW MODEL

随着现代医学成像技术的快速发展,医学影像分类已经成为重要的辅助诊疗需求。将文本领域中的词袋模型引入到图像领域,构建视觉词袋模型。为解决多义词和同义词问题,通过把词袋模型与PLSA主题模型结合,提出PLSA-BOA模型来解决传统词袋模型中的语义问题,这使得基于词袋模型的分类方法在精度上得到了进一步提高。实验结果表明,PLSA-BOW模型用于医学影像分类,具有较高的分类精度。

With the rapid development of modem medical imaging technology, medical image classification has become an important auxiliary diagnosis and treatment demand. In this paper we introduce the bag-of-words model in text field to image field, and build the model of visual bag-of-words model. To solve the problems of polysemous words and synonyms, we propose the PLSA-BOW model to solve the semantics problem in traditional bag-of-words model by combining the bag-of-words model with PLSA subject model. This makes the classification method based on bag-of-words model further improved in accuracy. Experimental results show that the PLSA-BOW model for medical image classification has higher classification accuracy.