Evaluation and analysis of an ordinal-based approach to video signature

In the current Internet age, there is a need for a set of good tools for copy detection purposes, as well as enhanced navigation, indexing and searching. We tackle this problem using an ordinal method. This paper presents an extension to our earlier work on an algorithm for generating a video signature based on an ordinal measure. We have achieved precision and recall of 0.935 and 0.966 respectively on our dataset, using a block size of 5 and a threshold of 0.004. This represents an increase in the discriminatory power of our algorithm.