Mobile visual product search services

• Main Author: Winata, Kaisarudin
• Other Authors: Yap Kim Hui
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/68391
• Institution: Nanyang Technological University

Due to the rapid development of technology, mobiles phone are considered to be an important gadget in everyone’s daily life and lifestyle. Nowadays, people using mobile phone to communicate with each other, playing games, search information in the internet, capture images and engage in social media network. Currently, there are numerous research and software development for image recognition but only limited to none image recognition software available for wine labels images in the wine stores or online. The purpose of this project is to assist mobile phone users to retrieve information of wine labels which they are interested in by easily capture the image of the wine labels by using their mobile phones’ cameras. This project will discuss several image recognition techniques and comparison of wine labels images under different conditions. Unwanted features in the surrounding wine label images and occlusions features by the users are the limitation when capturing wine labels images which produce a noise in images matching process. Thus, the matching accuracy results will be affected. Solution for the issue mentioned above is to introduce a technique called Geometric Verification (GV) in the image processing. This method advantage is it able to increased images recognition accuracy rate up to 10%. The minus point of this technique is extended processing time during features matching process. A set of experiments was conducted for image processing with and without Geometric Verification (GV) for comparison purpose. Finally, methods to improve matching accuracy when taking wine label images and recommendation for future work development will also be introduced.