近年来在对病人的消化道系统检查中,无线胶囊内诊镜(WCE)是一种最新技术,可以让医生直接观察到病人的病灶所在,但是对于消化道系统中的口腔、胃、小肠和大肠的WCE视频分类却是难点所在。相关研究中均采用通过人工标记的训练库的有监督学习方法。为了在WCE训练数据中获得高识别率,提出一种无监督学习方法,它利用融合颜色信息的尺度不变特征转换(SIFT)获取局部特征,再利用概率隐语义分析模型(pLSA)数据训练中进行语义内容分析。实验结果表明,在WCE图像分类中本方法与当前最新的监督分类方法一样可以获得高准确率。

Since Wireless Capsule Endoscopy (WCE) is a novel technology for recording the videos of the digestive tract of a patient,the problem of segmenting the WCE videos of the digestive tract into sub-images corresponding to the mouth,stomach,small intestine and large intestine regions is not well addressed in the literature.A few papers addressing this problem use a supervised learning approach that presumes availability of a large database of correctly labeled training samples.Considering the difficulties in procuring sizable WCE training data sets needed for achieving high classification accuracy,we introduce an unsupervised learning approach that employs Scale invariant feature transform (SIFT) with color information for extraction of local features and uses probabilistic latent semantic analysis (pLSA) model for data semantic analysis.Our results indicate that this method compares well in classification accuracy with the state-of-the-art supervised classification approach to WCE image classification.