Mobile product recognition service

• Main Author: Yuen, Pui Leng
• Other Authors: Yap Kim Hui
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/77398
• Institution: Nanyang Technological University

With the advent of digitalisation and Artificial Intelligence (AI), automatic recognition has become an increasingly important domain. This research project deals with the recognition of various brands of potato chips under varying environmental conditions of illumination, size resolution, angles and occlusion. In this report, we employed a simple yet robust preprocessing technique that first detects the packet of chips from an image and then corrects the illumination as needed. Therefore, this report focuses on studying the various algorithms. The processes and outcome of different feature extraction techniques were carefully studied: (1) SIFT feature and (2) Histogram of Gradient feature, both tested with K-Nearest Neighbour and Support Vector Machines as classifier respectively. An average hit rate of 98% was found for the best combination – namely the SIFT (in the case of all conditions namely; Angles, Sizes, Occlusion, Illumination and Distractors).