Publication Date


Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science


Within the past few years video usage has grown in a multi-fold fashion. One of the major reasons for this explosive video growth is the rising Internet bandwidth speeds. As of today, a significant human effort is needed to categorize these video data files. A successful automatic video classification method can substantially help to reduce the growing amount of cluttered video data on the Internet. This research project is based on finding a successful model for video classification. We have utilized various schemes of visual and audio data analysis methods to build a successful classification model. As far as the classification classes are concerned, we have handpicked News, Animation and Music video classes to carry out the experiments. A total number of 445 video files from all three classes were analyzed to build classification models based on Naïve Bayes and Support Vector Machine classifiers. In order to gather the final results we developed a “weighted voting - meta classifier” model. Our approach attained an average of 90% success rate among all three classification classes.