PEOPLE COUNTING SYSTEM IN CROWDED SCENES BASED ON FEATURE REGRESSION

While people counting has been improved significantly over the recent years, crowd scenes and perspective distortions remain particularly challenging and could deeply affect the count. To handle such problems, we propose a counting system based on measurements of interest points, where a perspective normalization and a crowd measure-informed density estimation are introduced into a single feature. Then, the correspondence between this feature and the number of persons is learned by Gaussian Process regression. Our approach has been experimentally validated showing more accurate results compared to other features-based methods.