Structured local binary Haar pattern for pixel-based graphics retrieval

Abstract

A feature, named structured local binary Haar pattern (SLBHP), is proposed for pixel-based graphics retrieval. The SLBHP is a hybrid of local binary pattern (LBP) and Haar wavelet. The SLBHP encodes the polarity rather than the magnitude of the difference between accumulated grey values of adjacent rectangles. The polarity relationships are then considered as a binary value as in LBP. Experimental results on graphics retrieval show that the discriminative power of SLBHP is superior to those of using edge points, Haar feature, and LBP even in noisy conditions.