Solving Object Detection and Localization in an Image.

Date of Submission

December 2017

Date of Award

Winter 12-12-2018

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science

Department

Electronics and Communication Sciences Unit (ECSU-Kolkata)

Supervisor

Mukherjee, Dipti Prasad (ECSU-Kolkata; ISI)

Abstract (Summary of the Work)

Object recognition in natural images using the database images taken under ideal lighting conditions has been a challenging problem in the field of computer vision. In this report, we solve the problem of identifying the products in the rack image of a grocery store. In the past few years this has been the interest of many computer vision researchers. We have the database images of different products available in a grocery store using which the products in a rack image need to be found. The problem can be divided into two major sub problems of matching and localization. We first use the matching from SIFT and try to improve the matching using a patch based matching algorithm. After matching, we are left with localization part and a product could be present at more than one location in the rack. We locate multiple instances of a product by clustering matched points using density based clustering methods. Although this method is vulnerable to outliers, we try to reject few of them by ignoring the less dense clusters. Finally we run this algorithm on our data set. We report good accuracy on our data set.

Comments

ProQuest Collection ID: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:28843249

Control Number

ISI-DISS-2017-359

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/6884

This document is currently not available here.

Share

COinS