How Do People Tag Picture? - A Pilot Study with Facebook Application

Acknowledgement First and foremost I would like give my best appreciation to my supervisor, Lexing Xie who offers a lot of support during the project development and composing this report. In the development process, she provides us a clear guidance of project direction and makes lots of useful suggestions towards problems encountered. Secondly, I would like show my thankfulness to my teammate Li Li. As teamwork, this project requires coordination along development. He is always friendly in our discussion and willing to listen to my opinions. Li Li's intelligence in design and analysis is really important in the successful of our project. Finally, I would like to thank all the participants that contribute the data collection of our project. Without your help, our project would not be possible. Abstract Photo tagging has been popularized by the increasing photo capture devices and diverse content sharing services. Considerable research topics have been achieved in this area: understanding motivations and usage of tags, the accuracy and value of machine generated tags, and using human brainpower to collect large scale photos tags via ESP games. In this project we try to figure out the differences between authors' tags and viewers' tags. To achieve this goal, a Facebook application called " Tag Expert " is designed and implemented to collect tags for Flickr photos by Facebook users. Then we designed several data collection methods and only choose one of them to evaluate due to the limited time and participants. The evaluation shows that GWAP principals could affect the correctness and efficiency of data collections for our application. Together with the experience in our developing process, the evaluation and untried data collection methods could be a very useful pilot study for future development in this area.