Content based image retrieval (CBIR), also known as query by image content(QBIC) helps users to retrieve the most similar and relevant images based on their contents. CBIR technologies provide an automatic way to find images in a large database in an efficient and effective way using one or more feature descriptors such as shapes, color, and text extracted to describe the distinctive features in the images. In this project, the features of the query image are compared with those of the images in the database in order to rank each indexed image according to its distance to the query. Before extracting the features from the query and compare it to the feature vector of the indexed images, two preprocessing filters applied on the images that helps in enhancing images and defining their objects.
The experiment in this project shows that the selected features provide an effective way in the retrieval process, but they couldn't be used as an accurate measure for classifying images due to their variability. The distance gives a good indication for classifying images and retrieving most relevant images.