The latest research done in the Deep-FIR project has just been published in MDPI Sensors, titled ‘The Best of Both Worlds: a Framework for Combining Degradation Prediction with High Performance Super-Resolution Networks’.
The paper details a novel approach where existing super-resolution models can be modified to utilise supplementary information in the form of degradation prediction in order to further improve performance and to better reflect real-world scenarios where degradation information needs to be estimated.
The paper may be found here. The code accompanying the paper is also available here, with models available for download here.
