As the increasing number of images emerge on the Internet, search image by keywords is not ideal or sufficient. Content-based image retrieval (CBIR) is one potential solution. However, CBIR researchers focused more on accuracy, efficiency and effectiveness of the matching algorithm, and there is the obvious gap between the users and the image matching algorithms in laboratory. Therefore a suitable user interface is essential to allow a user to interact with CBIR algorithm.We designed and developed a mobile App (Android) to allow users to draw and find their desired paintings. And two studies which are both related to user's drawing quality - a colour palette study and a visual memory study - were conducted on our mobile App. We aimed to explore the best elements to facilitate user to draw their search intention. In this talk, I will introduce (1) the aims of the study, (2) the methods used in each study, (3) the results we got so far, and (4) potential contributions of my phd work.
University of Nottingham School of Computer Science Nottingham, NG8 1BB
email: mrl@cs.nott.ac.uk