Interval type-2 fuzzy integral to improve the performance of edge detectors based on the gradient measure

In this paper we show the improvement of edge detectors based on the gradient measure using interval type-2 fuzzy logic. The improvement consists on the representation of uncertainty in image gradients and their aggregation using the interval type-2 fuzzy integral. The inclusion of uncertainty in gradients helps to find true edges that could be ignored with other methods. This method can be used to identify shapes in images with very variable contrast or in applications which need to find more edges in images that the classical methods.