realblur
Realblur is a term used in photography and computational imaging to describe motion blur that occurs in real-world scenes during exposure, as opposed to blur produced synthetically by applying a blur kernel to a sharp image. In research, RealBlur also refers to datasets consisting of real blurred images paired with corresponding sharp references, created to provide realistic benchmarks for deblurring algorithms.
The RealBlur datasets aim to capture the complexities of blur found in everyday photography, including camera
Two widely used subsets of the RealBlur collection are RealBlur-J and RealBlur-R, which originate from different
Impact and usage: RealBlur has become a standard resource for advancing real-world deblurring, encouraging models trained