A framework for audio feature extraction from videos for genre identification

Video genre identification is a process of determining the genre of the videos. Today, many private households, broadcasting and film industries have a large collection of videos in an unorganized manner. The main issue is that how to organise the large video collection into a meaningful subjects. It is a challenging task. To overcome this issue, the video genre classification concept is used. That is analysing the video, text and audio features. By considering only video or text features, the video files cannot be classified into predefined genres such as news, sports, cartoon etc., more accurately. The proposed work is to classify the video files into various genres using audio feature analysis. The audio features related to frequency domain and time domain is determined. Then the genre identification is performed on the audio features analysis by using rule-based classifier concept. The proposed system is efficient when audio features are taken into consideration for identification of genre of videos.