I. Introduction
In the present scenario, the number of news channels is increasing day by day and it has become very difficult to keep track of each and every news shown in a particular news channel. In many instances, the same news is conveyed in different channels with different viewpoints which further manipulates the interpretation made by a viewer. So, it is necessary to have a proper database of news broadcasted in different news channels. Different spatio-temporal features used by different channels further aggravates the problem of generalising a method of video sorting. Moreover, the sorting of such news videos accelerates the task of finding particular news among the lot. The news channels also display commercials in between the news videos. Analysis of the frequency of a particular type of commercial or a particular advertisement can also be done by proper sorting of the broadcast news videos. For the sorting of news videos, the preliminary step is a proper segmentation of the news videos. We have proposed an unsupervised shot segmentation algorithm in this paper as a stepping stone towards the sorting of news videos.