Mobile visual product search and recommendation
Nam, Freddy Weixiang
Date of Issue2016-05-21
School of Electrical and Electronic Engineering
Smart phones have evolved immensely throughout the years in its function. High resolution cameras, color displays and better graphics have produced good quality image and video through the phone. In additional, broadband wireless hot spots, GPS and wireless network technology (3G, 4G, etc) enables us to access the net easily. With all these made available, we can easily query for products or places through the net with the image captured by the user without using text based search. Mobile visual search application can be used to query images and the query result will be displayed on the camera page. In addition with augmented reality function, the query result is accurate and the response is fast. Augmented reality enables user to continuously scan the image till a result is displayed and therefore the time for searching is greatly reduced as compared to the traditional text based search. Further in the application, user will be enriched with product information and also recommendation and comparison of products. This in turn, can help to close the gap between the real and digital world.  Mobile Visual Search applications also pose a few challenges. 1. Query of images have to be fast in order to increase efficiency 2. Battery consumption while using the application 3. Robustness of object in terms of the shape of object and lighting condition at that point of time 4. Handling of constrained domains All these factors pose a need to further studies on the field of mobile visual search.
Final Year Project (FYP)
Nanyang Technological University