Synchronizing classroom documents

Abstract

Recently, automated process for classroom teaching and learning has become one of the main interested areas in multimedia authoring activities. One such application is the distance learning which aims to provide a quality assurance learning environment for teaching and training remotely. In this project, I study the synchronization of lecture slides with the audio clips which are captured from recent lectures. At the beginning, speech recognition is performed for the detection of phonemes in the audio clips. After that, phoneme sequence detection techniques are used to facilitate the detection of words in audio files. Finally, the Earth Mover's Distance (EMD) algorithm is modified for the matching of slots and slides. The results of this project give an evaluation on the performance of different parts. These include the evaluation on the correctness and accuracy of speech recognition, and the performance of the new modified EMD algorithm. Finally, we estimate the performance of the whole project after different parts are combined. The real situation is also simulated to give a completed evaluation of this project.