In this paper different variance filters for rejecting image regions that do not contain interesting object are tested. In our case the processed scenes have equally depth of focus, which makes difficult to distinguish objects from the background. In order to locate the object, the algorithm based on the sliding windows approach has been used. In case of using this type of algorithm a cascade of filters designed to reject windows that do not contain searched objects are applied. In this paper the authors put emphasis on elimination of redundant windows, from equally depth colour scenes, using various variance filters. Also a formula, based on the integral images, which can improve the efficiency of using directional variance filters, is proposed. All types of variance filters are tested and compared. © Springer International Publishing Switzerland 2015.
CITATION STYLE
Sarwas, G., & Skoneczny, S. (2015). Object Localization and Detection Using Variance Filter. In Advances in Intelligent Systems and Computing (Vol. 313 AISC, pp. 195–202). Springer Verlag. https://doi.org/10.1007/978-3-319-10662-5_24
Mendeley helps you to discover research relevant for your work.