Content based video retrieval based on approximate string matching

We present a query by example mechanism, to make queries from a video database. During a search for similarity matching, in a video database, two important points should be considered carefully, first the similarity between single frames and second, the similarity between the order of the frames in two video sequences. Based on these facts, we present a query by example video retrieval approach, in this paper. We introduce an algorithm for similarity matching, which finds the similarity between video sequences with different sizes. Moreover, the algorithm is able to find the similarity between a query video and a part of another video sequence. During this process a confidence rate will be considered, which makes the approximate similarity matching possible.