Visual features used for search of visual material usually have computationally complex similarity functions. Therefore for large databases to get real time response for queries by examples is necessary avoiding their full search. In this paper we show efficiency of selected techniques for accelerating visual object retrieval. They belong to three independent groups: filtering, partial similarity computing, and tree based data structures. We show on description examples of motion trajectory, face recognition, and distributed color image temperature that different types of visual features require different accelerating techniques. © Springer-Verlag 2004.
CITATION STYLE
Galinski, G., Wnukowicz, K., & Skarbek, W. (2004). Accelerating multimedia search by visual features. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3211, 729–736. https://doi.org/10.1007/978-3-540-30125-7_90
Mendeley helps you to discover research relevant for your work.