Implementation of Local Binary Pattern in Feature Recognition and Novel Approach to Classify the Images Based on Texture Features

Abstract

A novel approach for defining the features by the Haar classifier based on algorithm Local Binary Pattern. In this system, we have used haar feature-based cascade classifier for object detection. Local Binary Pattern has been used due to their computation simplicity and stoutness towards variation in illumination, which considers pixels of a picture by thresholding the space of each pixel and contemplates about the outcome as a binary number. The main purpose of this paper is to develop an application, which classifies the images, train them and extract features. The test data is then matched with the stored patterns using local binary pattern (LBP) algorithm. The aim of the application and research is to enhance the robustness and computation power by inducing parallel programming during the training phase.