Publication details

Sub-image Searching Through Intersection of Local Descriptors


HOMOLA Tomáš DOHNAL Vlastislav ZEZULA Pavel

Year of publication 2010
Type Article in Proceedings
Conference 3rd International Conference on Similarity Search and Applications (SISAP 2010)
MU Faculty or unit

Faculty of Informatics

Field Informatics
Keywords sub image retrieval
Description The complexity of search in current business intelligence systems, academic research, or even the home audiovisual databases grows up rapidly. Users require searching by the content of their data. For example, the user sees a cathedral while watching a movie and by taking a snapshot, his or her private collection of holiday photos can be searched for images containing that cathedral. In practice, it is not sufficient to store data and search in it by exact match but rather by means of similarity, i.e. retrieving data items similar to a query item. Similarity searching is especially requested in multimedia databases, digital right management systems, computer aided diagnosis, but also in natural sciences and psychology. In these fields, theoretical primal background for querying by similarity is already defined. In this demonstration, we focus on similarity searching from the perspective of identifying and locating an object depicted in a query image in a collection of images.
Related projects: