الفهرس 
acknowledgementOnly 14 pages are availabe for public view
 |  | from | 83 |  |  |
| | | from | 83 | | |
Abstract
Latest image technology improvements along with the internet and social networks growth have led to huge amount of digital images during the recent decades. Various methods, algorithms and systems have been proposed addressing image storage, indexing, retrieving and management problems. Such studies revealed the indexing and retrieval concepts, which have further evolved to Content-Based Image Retrieval (CBIR). CBIR systems often analyze image content via the so-called low-level features for indexing and retrieval, such as color, texture and shape or combinations of two algorithms. However, such single or combinations increase the feature extraction processing time and memory requirements as well as the retrieval complexity. It is worth mentioning that performance improvements of indexing and retrieval play an important role for providing advanced CBIR or search engines services on every hardware platform.
The main aim of the study is to improve the overall CBIR system performance in different search engine hardware platforms having different technical capabilities and conditions.
In this thesis, we propose novel techniques for improving the overall performance of CBIR. We define general CBIR challenges as memory and disk space requirements, computational complexity, semantic retrieval performance and usability.
A novel system for feature extraction and selection is introduced for enhancing semantic image retrieval results, decreasing retrieval process complexity, and improving the overall system usability for end-users of CBIR systems. The research proposes two algorithms: Fast Scale Invariant Feature Transform (F-SIFT) algorithm and Ontology-Tagging algorithm.
Experimental studies also show that the proposed F-SIFT algorithm is a fast and accurate feature extraction algorithm. We also studied the effects of image transformation on feature extraction algorithms and on semantic retrieval performance to utilize the proposed algorithm and framework scheme.