Content-Based Image Retrieval: State-of-the-Art and Challenges

Images are an increasingly important class of data, especially as computers become more usable with greater memory and communication capacities. Development in digital photography has led to a huge collection of still images that are stored in digital format. As the demand for digital images increases, the need to store and retrieve images in an efficient manner arises. Therefore, the field of content-based image retrieval has emerged as an important research area in computer vision and image processing. The key issue in image retrieval is how to match two images according to computationally extracted features. Due to applicability of content based image retrieval in various domains, researchers are now merging fields such as machine learning and computer vision with image processing. This merging of fields provides us an opportunity to find the solutions of the issues such as image annotation, semantic gap, dimensionality reduction and so on. Keywords-content-based image retrieval; feature extraction; similarity measures; relevance feedback