In a typical content-based video retrieval system, the user submits a query and will then expect the system to locate all similar data in the corresponding video database. Content-based means representing the image content by means of low level visual features that are extracted through image processing techniques. The low level features include colour distributions, textures, shape and motion, in the case of video. In this thesis, a novel content-based video retrieval system using shape features is proposed. To provide indices, first all the shots in a video sequence are extracted. Then corners of all frames in each video shot are detected applying proposed curvature scale space corner detector using different scale of smoothing. As a user interface, a proposed fast active contour model is used to specify one object of interest as a query in one of the frames of each video shot. Afterwards, closest corners of query object to the final snake in this frame are extracted and tracked forward and backward through the whole of that shot using proposed multiple-match tracker. This tracker, which does not make any important assumptions or use any motion models, can retrieve the query in any video sequence even when there is non-smooth and unconstrained motion. By tracking the corners of query object forward and backward, the positions of similar objects in each video frame are determined. Two methods are considered for demonstrating the query and its similar objects to the user. Experiments have been carried out on a wide range of real video databases. All the results confirm that the proposed method which exploits frame corners is more efficient and more generally applicable.