Journal article Open Access
The massive volume of data stored in computer files and databases is rapidly increasing. Users of these data, on the other hand, demand more complex information from databases. The video data have exponential growth towards accessing and storing. The vital problem associated to video data is efficient, qualitative and fast accessing. We talk about how video pictures are clustered. We presume video clips have been divided into shots, each of which is denoted by a collection of key frames. As a result, video clustering is limited to still key frame pictures. In amble database finding the qualified data set (clusters) is quite time-taking job. The video data mining relate to multi–lingual text, numeric, image, video, audio, graphical, temporal, relational and categorical data. It may be any kind of information medium that can be represented, processed, stored, fast accessing or summarization of clusters are required due to which significant frame-set is formed. Due to sampling error and test reliability in video, substantial changes of more than one frame are predicted. The goal of this article is to show how to employ a familiar and easy nonparametric statistical approach (chi-square) to select eligible data/framesets for analysis. The chi-square model illustrated here is a straightforward, sensible, fast, reduce saddle, and easiest method. Skimming/ Summarization and clipping technique are further enhanced by this technique along with video database maintenance technique from simple descriptors to a complex description schemes like spatial and temporal or high dimensional indexing.