AlikE Content Detection in Image and Video
Shibi Thambi k, Vidya R Menon, S Sreerag · International Journal For Science Technology And Engineering · 2015
This paper recognizes the problem of detecting alike content from images and video for content based retrieval applications. With the development of digital multimedia data types and available bandwidth there increase a demand of image retrieval systems and the users shift from text based retrieval systems to content based retrieval systems. Selection of extracted features play an important role in content based retrieval systems. These features are used for indexing, selecting and ranking according to the user. Good features selection also reduces the time and space costs of the retrieval process. Content based image retrieval (CBIR) is a technique in which it uses visual contents to search images from the large image or video databases. Alike features can be derived from a set of feature extraction methods like SIFT and MSER. Here the features generated are then combined together to form feature code book. Both mid-level feature description techniques and multi class support vector machine is used as classifiers for detecting purposes. Analysis is done using four different image classes’ viz. cars, flowers, buildings and human faces.