Performance of media based information leakage protection

There are many occasions that an organization needs to distribute important data to only authorized agents. However, there might happen illegal interceptions of the data by unauthorized agents. Commonly used methods of protecting information from unauthorized accesses have been based on text, but there have also appeared media based approaches in the literature. In this paper, we model and implement several combinations of the authentication systems based on the methods of using text, image and video, and evaluate their performances. In our practice, we developed 13 refined combinations of the testing models and implemented them in Java with Visual Studio as the front-end and a SQL server as the back-end. Performance analysis is done comparatively in the aspects of information protection rate and system access time delay. The experimental results revealed that the model with the combination of a text and 1 or 2 images with fine-grained access yielded the best performance with very low overheads.