Vyhledávání v kolekci obrazků pomocí komprese

Abstract

Searching and comparing images is one of actual topic in information retrieval domain. Because of a huge amount of images on the internet and local storages, annotation based approach is decreasing. In my thesis, I have been working on more effective methods such as content based image retrieval. Because of the most of the images are in compressed format I am working with information retrieval on compressed data by vector quantization. The main goal of this thesis are retrieving images to benchmark and implementing system for indexation and searching images. At the end I am describing achieved results and I suggesting some ideas to improve the system.