Develop a browsing system for tagged images

Abstract

Tag is becoming a key supplementation to the information of the social media sharing items. With the rise of social media sharing, tag attracts more and more attentions from researchers. A lot of concepts are proposed to facilitate tags’ representativeness, like image tags’ visual-representativeness. This project is undertaken to develop a Java application to map and visualize the visual-representative relationship among the tags of images from Flickr. It may help to uncover the human behavioral pattern of annotating images by demonstrating the visual-representativeness and the correlation of the tags. The interface of the application has been implemented with Java Swing component and open source APIs TouchGraph. Though there is a flaw on the response time, the application design has well accomplished all the objectives of the final year project. Discussing the successes and flaws of the project, a conclusion reflects on progress and proposes further opportunities for new development.