Speaker Diarization in Personal Video Recordings Based on LDA and User Feedback

In this paper, we present the speaker diarization system which is used in personal video recordings. Speaker diarization begins by the extraction of relevant features from the input signal. Features are measurable characteristics which are important to the distinction between different classes. They should have low inter-class similarity and also low intra-class variability. So, LDA is used to cope intra-speaker variability. We demonstrate improvement of the performance over the baseline system based on LDA. This paper also reports the performance of the speaker diarization system. A new approach, called the user feedback has been developed here to improve the performance of the speaker diarization system. 