Mobile visual product search

• Main Author: Zaw, Kaung Myat
• Other Authors: Yap Kim Hui
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/74631
• Institution: Nanyang Technological University

Mobile devices nowadays have developed into effective video and image processing tools fitted with high resolution displays, cameras with amazing functions and powerful graphics. They have also been equipped with features and applications that allow users to perform almost every task in a touch of a finger and that is why mobile devices have become a necessary component in people’s everyday lives. With the advancement in technical capabilities of mobile devices, the increase in development of new classes of applications also takes place. Augmented Reality (AR) and Virtual Reality (VR) have currently become a hot area of research in the computer vision processing field. More and more mobile entertainments and lifestyle applications are emerging in which the full potential of smarter mobile phone cameras is utilised. This enables a new form of application where users can make use of phone cameras as input for objects that they want to query and search information about. This is similar to QR Code scanners where users no longer need to go into specific search engine platform and type long query texts to search for information. Such kind of application can be of help for product identification and comparison while shopping and browsing real-time information for any movies, albums, books, artworks or architectures. Mobile Visual Search (MVS) allows users to simply take a snap of a product using mobile phone camera and look up necessary information of it from the internet, such as price, specifications, availability and promotions. Since the scope of the project is to implement a system for Mobile Visual Product Search, as the very first step, we needed to determine a product scope to focus on. Beverage products have large varieties of choices and they come in different shapes and designs. Thus, choosing on this product scope allows us to determine the system’s performance under beverages from same brand but different types as well as different shapes. Due to the wide availability, we could gather various kinds of beverage products needed for our project and we tested out the consistency and robustness of our MVS algorithm.