| Summary: | Mobile Visual Search is a trending image search technology leveraged on mobile phones. One can take an image on the mobile phone, process the image, and send it to an online server to find the best matching result. The result is then sent back to the mobile phone for applications such as landmark recognition, printed media information retrieving and so on. With the rapid growth of the Internet, a large number of food recommendation websites have sprung up, and some of them have acquired a considerable number of users; moreover, food sharing has always been a favorite topic on various social media. However, currently there are very few applications out in the market that can seamlessly integrate the restaurant recommendation function with social media attributes. In this project, an Android mobile application was built from scratch. It aims to provide the restaurant recommendation service to users, meanwhile enable users to have social interactions with their friends.The first and foremost function of this application - the restaurant search function was realized by Mobile Visual Search. Performance test has shown that the underlying image search engine was able to achieve an average of 87% recognition rate with an average of 1.77 seconds recognition time.A friend-based review system was also designed in this application to provide a reliable and interactive restaurant recommendation function. Two kinds of browsing lists were provided to instantly update users with the most recent or most popular restaurants among their friend circles. The application itself has proved to be stable and efficient during various use-case tests.
|
|---|