Content-Based Image Retrieval (CBIR), also known as Image Content Query (QBIC) and content-based visual information retrieval (CBVIR), is the application of computer vision techniques to the image retrieval problem, ie , Images in large databases (see this survey for a recent scientific view of the CBIR field). Content-based image retrieval is opposed to approaches based on traditional concepts (see concept-based indexing).
"Content-based" means that the search parses the content of the image instead of the metadata, such as keywords, tags, or descriptions associated with the image. The term "content" in this context may refer to colors, shapes, textures or any other information that may be derived from the image itself. CBIR is desirable because searches that rely solely on metadata depend on annotation quality and integrity. The fact that humans manually annotate images by entering keywords or metadata into a large database can take a long time and do not capture the desired keywords to describe the image. The evaluation of the effectiveness of the search of images of keywords is subjective and has not been well defined. In the same sense, CBIR systems have similar challenges in defining success.
The term "content-based image retrieval" appears to have originated in 1992 when it was used by T. Kato to describe experiments in automatic image retrieval of a database, based on the colors and shapes present. Since then, the term has been used to describe the process of retrieving desired images from a large collection based on syntactic image features. The techniques, tools and algorithms used come from fields such as statistics, pattern recognition, signal processing and computer vision.
The first CBIR trading system was developed by IBM and was called QBIC (Query by Image Content). Recent approaches based on networks and graphics have presented a simple and attractive alternative to existing methods.