In solving what tasks can CBIR aid?
by Sergey Egorov
Here the most actual applications of CBIR are presented:
- detection of objectionable content (porno, erotics and so on);
- person identification (by means of face detection and recognition techniques), matching with celebrities (like Picitup Celebrity Match);
- criminal identification (fingerprint, face, iris);
- search by image sample, outline and sketch (by colour, set of colours, texture, shape, spatial layout and other features);
- search of similar parts and fragments in images (sub-image retrieval);
- images duplicates detection, clusterization (with similar images in each cluster) and classification of images collection;
- organization of photo archives and multimedia encyclopedias;
- visual shopping (examples: Like.com, Picitup, Etsy, Visuvi);
- geographic information systems, processing of different maps, satellite and aerial photos;
- in fashion and textile industry (texture analysis);
- in absolutely different spheres (stenography, astronomy, ecology and environment and etc.);
- online applications for art and art history (examples: QBIC Project and Wang Project);
- medical image archives and processing, pathology detection and aid in diagnostics (examples: IRMA Project, Visuvi);
- trademarks and logos databases and revealing the infringement of copyright (like PixID);
- the similar applications for mobile phones (like Picollator.mobi) and iPhones (like Apple);
- the similar applications for video retrieval (search of similar persons by faces and similar scenes by fragment, sample, sketch, spatial arrangement of colours and regions and etc.), example – TV news processing;
- document image retrieval – revealing the text in images, signatures, stamps, handwritings and words matching, search by text query without any character recognition.
Of course, it is far from being the complete applications list!
Written by Sergey Egorov
No tags for this post.