Herzog, OttheinJacobs, ArneArneJacobs2020-03-092020-03-092010-11-17https://media.suub.uni-bremen.de/handle/elib/50Interest point detectors are a common processing step in low-level computer vision. Current interest point detectors exhibit invariance against rotation, scaling, and affine twodimensional transformations. The thesis presents a new algorithm for interest point detection that is invariant against a larger set of transformations, i.e., deformations.deinfo:eu-repo/semantics/openAccessImage processingcomputer visionsalient points000Dissertationurn:nbn:de:gbv:46-00101784-17