Visual product search in mobile business and advertisement

• Main Author: Cao, Wei Dong
• Other Authors: Yap Kim Hui
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/63793
• Institution: Nanyang Technological University

In this report, the visual search application technology that allows the clients to upload images and information related to the products to the server and database that can help the users search for the queries by capturing an image using mobile phone is introduced. We propose a system architecture where by using this application, a client can upload and transmit the image and details of a book to the server and database. The user can capture a photo of a book and the image will be transmitted to the server. At the server, the query image features are extracted and used in the recognition from databases. Once identified, the result will be returned and displayed on the screen of the mobile device. In order to demonstrate this capability, we have designed and developed a Visual Search application. The mobile application designed by past year FYP students requires a web application to provide the server and database for the backstage and also for more platforms to operate. The web application can also be used for the clients to upload the images and data to the server and database as the source for the books. To further enhance the application contents and user interface, we enhance the application to allow the multiple data or multiple images to be uploaded together. We also enhance the integration between the web application and mobile application. Experimental results show that the web transmission system can upload single and multiple images to the server with the overall performance duration of less than 200ms for single image. The results can be seen in the server and through the web application. The multiple data in the correct form can also be uploaded to the database with 100% accuracy.