Experiment with reduced feature vector in CBIR system with relevance feedback
Goran J. Zajic, Nenad S. Kojic, Nikola B. Reljin, Branimir D. Reljin · 2008
A CBIR system with relevance feedback is described. System uses feature vectors (FVs) with only 24 components describing color, line directions and texture. The reduction of FV dimension is based on the statistics of global image features. The proposed system was tested over Corel IK dataset and results are compared to some other CBIR systems with or without FV reduction.