Fast HDR image upscaling using locally adapted linear filters

摘要: A new method for upscaling high dynamic range (HDR) images is introduced in this paper. Overshooting artifact the common problem when using linear filters such as bicubic interpolation. This visually more noticeable while working on HDR where there exist transitions from dark to bright. Our proposed capable of handling these artifacts by computing a simple gradient map which enables the filter be locally adapted image content. adaptation consists of first , clustering pixels into regions with similar edge structures and second learning the shape length our symmetric each pixel groups. can implemented in a separable fashion perfectly fits hardware implementations. experimental results show that training filter with effectively reduce overshooting improve upon visual quality existing linear approaches.