Study On Haar Classsifier For Machine Learning System That Can Detect Specified Object In Vision System

Chan, Horng Tat (2010) Study On Haar Classsifier For Machine Learning System That Can Detect Specified Object In Vision System. Project Report. UTeM, Melaka, Malaysia. (Submitted)

	PDF (24 pages) Study_On_Haar_Classsifier_For_Machine_Learning_System_That_Can_Detect_Specified_Object_In_Vision_System_-_24_pages.pdf - Submitted Version Download (338kB)
	PDF (Full Text) Study_On_Haar_Classsifier_For_Machine_Learning_System_That_Can_Detect_Specified_Object_In_Vision_System_-_Full_Text.pdf - Submitted Version Restricted to Registered users only Download (1MB)

Abstract

Nowadays, visual system is widely applied in the manufacturing process. However, implementation of visual system is highly cost due to expensive camera and software is needed. Therefore, a low cost object detection system with webcam is shown. The main purpose on this project is to create a Haar classifier that able to detect specified object. Haar classifier is one type of detection system application that apply boosted rejection cascade. It has applied decision tree with boosting theory in order to have better performance and high percentage of detection. 2 main categories that related to this topic are machine learning and machine vision. Haar classifier can be consider as a software that able to learn and differentiate the difference in a still image and it also can say as have vision to “see” and differentiate the images. The Haar classifier that is used in this project is based on Viola and Jones theory. Objects used for detection system are “T” shape object and transparent mouse to differential the effect of complexity object in the performance of system.

Actions (login required)

View Item