Build a Content Based Image Retrieval System of Endoscope Image

With the growing available multimedia digital image database, the question comes to how to design a good image searching engine or image retrieval system that is efficient in finding the images that users require. For medical educators and clinical researchers, with a handy image retrieval system means the saving in the time and money of data collection. It brings the value of GIFT (GNU Image Finding Tool), a open source content based image retrieval (CBIR) system. Still, free and open source give the system more flexibilities in practice. All of the above construct strong incentives to build a CBIR system and apply it to medical image. In this research, we use the open source CBIR system, GIFT, as a basic of medical image retrieval system of the endoscope images inside Interactive Clinical Image Bank of Taipei Medical University Hospital. By using open source tool, we establish a CBIR system with low cost and short time. We choose Gastric Ulcer and Gastric Cancer images to do evaluation of the medGIFT system and comes a good result which means the combination of GIFT and medical image database has great potential on medical education or even diagnostic support.