作者: A Anil Kumar , N Narendra , P Balamuralidhar , M Girish Chandra
DOI: 10.23919/EUSIPCO.2018.8553237
关键词:
摘要: This paper considers the problem of super-resolution (SR) image reconstruction from a set totally aliased low resolution (LR) images with different unknown sub-pixel offsets. By assuming translational motion model, linear compact representation between LR spectrums and SR spectrum, based on multi-coset sampling is provided. Based this we formulate joint estimation shifts spectrum as dictionary learning alternating minimization approach employed to solve estimation. Two approaches for obtaining image; one estimated another estimate are described. The significant advantage proposed smaller matrix sizes be handled during computation; typically order number enhancement factors, completely independent actual dimensions images, hence requiring significantly lesser resources than current state art approaches. Brief simulation results also provided demonstrate efficacy approach.