Tracking moving people with hausdorff distance-based matching using an adjustable-size search window

Abstract

Tracking moving objects based on a Hausdorff measure approach has become popular for computer vision applications [4, 5, 15, 16]. The main drawback of this approach is its computational complexity. Without any restrictions in the search space, it not to be suitable for real-time applications. In this paper, a new technique is described to reduce the computational cost based on the local search size estimate. The proposed method lets dynamical size fitting of the local search window using the knowledge and measurements provided by Hausdorff matching together with a predictive process. Experimental results with real world image sequences are provided to illustrate the performance and the considerable reduction of computational costs with respect to previous approaches and the results are analysed