Multiresolution LBP for lace images classification

Abstract

The lace images are particularly difficult to be analyzed in digital form using classical image processing and machine vision techniques. The major reasons of this difficulty emerge from their fine complex structure and pattern. In this work, the major approaches implied for the evaluationsof lace images are Image Histogram and Local Binary Patterns (LBP). Two variants of LBP are mainly evaluated. The accuracy obtained of each descriptor via the k-NN classifier with different distances. Experimental results indicate that the basic LBP generates the best performance 97% under neutral variation