Cattle Recognition using Fuzzy Speeded Up Robust Features (F-SURF)

Abstract

This paper presents a rotation-invariant detector and descriptor, using fuzzy-SURF (SpeededUp Robust Features). Fuzzy SURF helps to increase the schemes with respect to continuous, distinctiveness, and vigorous, which are comparatively much faster. Muzzle (viz. Nose) patterns are the asymmetrical features of the skin of cattle on its surface. The muzzle pattern can be considered as a biometric identifier for cattle. Image convolutions are done by relying on integral images; by building on the strengths of the leading detectors and descriptors (especially a detector based on Hessian matrix and a distribution descriptor); and by matching with fuzzy similarity measure. This leads to a combination of novel detection, description, and matching steps. The paper encircles a detailed description of the detector and descriptor and then explores the effects of the most important parameters