A Comparative Study on Various Techniques for Image Retrieval

In the digital world, there is a rapid increase in data that is being generated everyday. Obviously, the image data growth is also more. So from a large database containing images it is really hard to mine retrieve images that are relevant for the query. Image Retrieval is a significant research area in the domain of image processing. It contains features for extraction such as shape, texture, colour etc., for image comparison. Currently, the focus of research is in semantic gap reduction between high level image semantics and low level visual feature.. In this paper, a comparative study on various techniques for retrieval of images is studied. Multiple feature extraction can be done by combining various methods. Based on the extracted feature any of the classification techniques can be applied which will significantly reduce retrieval time and search space. Once when this is done, for the respective relevant images relevance feedback algorithm is applied which in turn provides the user intention for resultant images to the system as this increases the classification accuracy. This is done by getting the feedback from the user which in turns decreases the semantic gap.