Content-based blur image retrieval using quaternion approach and frequency adder LBP

Abstract

The paper presents a content based image retrieval scheme based on feature extraction and weighing. Features are extracted using frequency adder based local binary pattern and blur detection metric which are then optimally combined using a weighing scheme. Simulations are performed on modified Wang and KTH-TIPS databases, which include images from four different classes of blur respectively. Comparison of simulation results with the state-of-the-art techniques show better retrieval precision and recall values for proposed technique.