Content-based Image Retrieval for Medical Application

Image retrieval is an image searching system that uses certain characteristics or context in an image. In a medical field, image retrieval system has been used to provide the needed correct images to the physicians while the diagnosis and treatment process is being conducted. The earlier image retrieval system was a text-based image retrieval system (TBIRS) which using keyword for the image context and it requires human’s help to manually make text annotation on the images. This system is still being widely used in the hospitals nowadays. The text annotation process is a laborious task especially when dealing with a huge database and is prone to human errors. To overcome the aforementioned issues, the approach of a content-based image retrieval system (CBIRS) with automatic indexing using visual features such as color, shape and texture is introduced. However, it is well known fact that a specific algorithm in CBIRS is only applicable to one specific modality and body part. In addition, CBIRS ignores a human perception of an image semantic that leads to a semantic gap problem. Hence, hybrid based image retrieval system (HBIRS) which combines both TBIRS and CBIRS has been proposed to reduce the semantic gap. A framework of efficient retrieval system which is HBIRS are presented, however the development of HBIRS could not be implemented due to the weaknesses in TBIRS. Hence, this study only considers both TBIRS and CBIRS for medical applications, and a TBIRS prototype tested using various X-Ray images has been proposed.