A fuzzy approach to image filtering in diving video processing

Abstract

The aim of this paper is to compare two different types of filter for the diving video processing. The two filters are a boolean filter and a fuzzy filter. These filters are applied for improving the diving video analysis aimed to introduce quantitative tools and diving performance measurement and therefore to improve training. The aim of the filter is to identify the athlete in the video to further process the data in order to obtain information as the position of the barycentre and the rotational inertia. The application of the fuzzy logic to the filter is aimed to reduce the errors in identify the athlete when the lightening changes and consequently in the traditional filter it can be excluded.