Abstract:The aim of Multi-frame super resolution (MFSR) is to estimate a high resolution image from a set of low resolution images. This task is highly ill-posed and computationally costly. A new MFSR forward model is proposed that reformulate the MFSR problem into a problem of multi-frame blind deblurring (MFDB) which is easier to be addressed than the other. MFBD problem can be sufficiently solved by attacking the resulting subproblems of alternating minimization using the alternating direction method of multipliers (ADMM). Experiments can be performed on synthetic and real images and can show the quality of the image and high speed in execution.