基于QT的胶囊内窥镜图片分析系统

摘要

In light of heavy workload of going over all the wireless capsule endoscopy (WCE) images, we present a QT-based implementation scheme for WCE image analysis system. Based on classifying the mass WCE images, the system screens in turn the over-dark images and the duplicate images, and recognises the tumour lesions amongst the above filtered results. According to the analysis on the experimental result , the screening manners of either screening by grade or manually inputting the screening proportion have filtered out most of invalid and redundant image data, it greatly improves the film-reading speed of doctors while ensures their diagnosis quality.%针对现有遍历胶囊内窥镜海量图片的繁重工作,提出一种基于QT的胶囊内窥镜图片分析系统的实现方案.该系统在对海量图片进行分类的基础上依次进行过暗图片筛选、重复图片筛选,并在以上的筛选结果中进行肿物病变的识别.根据实验结果分析,依等级筛选或手动输入筛选比例的筛选方式筛除了大部分的无效和冗余图片数据,极大地提高了医生的阅片速度,同时保证了诊断质量.