This paper describes a novel approach to capsular endoscopy that takes advantage of existing wireless capsule endoscopy (WCE) and overcomes some of its important limits. The basic and essential task concerning the integration of SVM classifier is listed and discussed in details, as well as the new features required to improve its diagnostic capability. In addition, a large scale demonstrator has been evaluated and tested. In vitro experimental results were encouraging and show correct classification rate of intestinal polyps of approximately 93.7%. The work contains detailed statistics about the detection rate and the computing complexity.