Content Based Video retrieval is another facet of Content Based Retrieval systems. It is to be noted that videos are far more complex than images as they are a sequence of changing images and audio tracks. Here the content of the videos itself is matched with the query of the user. Video retrieval consists of a series of steps such as segmentation, frame selection, feature extraction and similarity measure. This paper is based on the fact that a video is a number of frames i.e. sequence of individual images which are moved at a certain speed. Therefore video is a sequence of moving images and same procedures used of Content based Image Retrieval for feature extraction can be applied to the video retrieval as well. The paper presents the approach called Color Difference Histogram which exploits the perceptually uniform color differences between two points under different background in accordance with the color and the edge orientations. As the system is based on the content of the videos, the query used here is a snap shot taken from the video recorder.