Beta 1

Title Context-aware photo gallery for mobile phone
Author Luniewski, Maciej Adam
Supervisor Larsen, Jakob Eg (Intelligent Signal Processing, Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
Institution Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
Thesis level Master's thesis
Year 2009
Abstract The constant development of the mobile devices, especially mobile phones, implies the increase in the image quality of taken photos. As the result, the size of personal image collections is very likely to expand in the coming years. Existing solutions, although sufficient today, will become inappropriate for management of collections of thousands of files. The suggestion was to implement a new application, which is based on the tagging concept. In order to address the problem with the image annotation, the concept of context awareness was exploited. It used various contextual data, available through the numerous sensors included in modern mobile phones. The scope of the project includes a theoretical research, a users survey, experiments with contexts and a prototype’s design, implementation and evaluation. The survey’s results revealed that time, place, events, and people belong to the most desired viewing and searching options. The field experiments have indicated that the use of Cell ID for location discovery is sufficient for the purpose of content management. The tests that followed showed that the accuracy of the determined place may be problematic in photo management. The last experiment indicated that GPS is the most battery consuming context, WLan is in second place, and Bluetooth is last. Based on the findings from the survey and the experiments the functional prototype has been implemented. The novelty features in mobile solution are views by place/city, event, days, weeks and months. Additionally, the prototype serves two searching functions: by various criteria and for the similar images. The evaluation confirmed a very good performance of the new approach and its appropriateness for the large image collections on mobile phone.
Imprint Technical University of Denmark (DTU) : Kgs. Lyngby, Denmark
Series IMM-M.Sc.-2009-20
Original PDF ep09_20.pdf (5.76 MB)
Admin Creation date: 2009-03-31    Update date: 2010-10-28    Source: dtu    ID: 240852    Original MXD