Image and audio sequence visualization and interaction mechanisms for structured video browsing and editing

We discuss extensions to our video browsing tool described by Hjelsvold et al. (see Handbook of Internet and multimedia systems and applications, 1998). We propose to include audio related informative visual content into our tool. Hence, we suggest representing the audio track of the video by its spectrogram image and pitch curve to enhance the video and audio related information available to the user. This representation also facilitates the correction of automatically computed audio event boundaries and introduction of speaker segments. We also provide a real-time approach for segmenting audio into events, namely, silence, speech and non-speech, to further enhance the audio information space.