The impact of technological development and digitalization in the field of communications and information processing, greatly influenced the realm of medical science, especially in the computer-aided diagnosis. The use of digital radiology in the past decade greatly accelerates the process of diagnosis and reduces the cost of storing large amounts of data. In addition, the digital radiology opened up new opportunities for research on large amounts of data. Medical images play a central role in patient diagnosis, therapy, surgical planning, medical reference, and training. With the advent of digital imaging modalities, as well as images digitized from conventional devices, collections of medical images are increasingly being held in digital form. It becomes increasingly expensive to manually annotate medical images. Consequently, automatic medical image annotation becomes important. In addition to medical image processing, an increasing need for processing large quantities of medical video clips and their use in diagnosis has increased in recent years. Due to the large number of the video frames without text information, content-based medical image retrieval (CBMIR) has received increased attention. The demand for automatically annotating and retrieving medical images and videos is growing faster than ever. Technological developments have enabled the use of small cameras that capture video clips in high resolution, allowing for a minimally invasive or even non-invasive procedure in patients. The best example of this use of modern technologies is the use of cameras for endoscopic examinations. These devices can capture long video sequences, which can then be processed and used in the detection of early stages of tumors and internal bleedings. Recently, the use of capsule endoscopy has become a way to record images of the digestive tract. The capsule is the size and shape of a pill and contains a tiny camera. After a patient swallows the capsule, it takes pictures of the inside of the gastrointestinal tract. The capsule can record a few hours of videos and the processing time of such a video is extremely large, making the analysis extremely difficult. This usually leads to failure in the detection od regions of interest. Manually processing the database that contains a large quantity of medical video clips has become very time consuming task with high potential for error in the interpretation of the content. Therefore, it is extremely important to explore the possibility of automatic detection of areas showing pathological changes or indicate the possibility that pathological changes occur. This would allow for faster and more reliable diagnosis. Automatic identification of regions of interest highly depends on the image retrieval and image classification.