2020/11/16 by Santiago López-Tapia, Santiago Lopez-Tapia, Nicolás Pérez de la Blanca +1 · 11 citations
Computer Science · Engineering · Physics and Astronomy · #Advanced Image Processing Techniques #Cascade #Convolutional neural network #Deblurring #Digital Holography and Microscopy #Focus (optics) #Image and Video Quality Assessment #Image restoration #Kernel (algebra) #Kernel density estimation #Pattern recognition (psychology) #Residual #Robustness (evolution) #cs.CV #cs.LG #eess.IV
paper · pdf · doi:10.1109/tip.2021.3074821
published in IEEE Transactions on Image Processing 30, 4747-4759 (Institute of Electrical and Electronics Engineers) · 12 pages and 11 figures (8 figures and 3 tables)
arxiv created 2020/11/16 · openalex created_date 2020/11/23 · openalex publication_date 2021/01/01 · arxiv updated 2021/05/19 · openalex updated_date 2026/08/05
Single Image Super-Resolution (SISR) is one of the low-level computer vision problems that has received increased attention in the last few years. Current approaches are primarily based on harnessing the power of deep learning models and optimization techniques to reverse the degradation model. Owing to its hardness, isotropic blurring or Gaussians with small anisotropic deformations have been mainly considered. Here, we widen this scenario by including large non-Gaussian blurs that arise in real camera movements. Our approach leverages the degradation model and proposes a new formulation of the Convolutional Neural Network (CNN) cascade model, where each network sub-module is constrained to solve a specific degradation: deblurring or upsampling. A new densely connected CNN-architecture is proposed where the output of each sub-module is restricted using some external knowledge to focus it on its specific task. As far we know, this use of domain-knowledge to module-level is a novelty in SISR. To fit the finest model, a final sub-module takes care of the residual errors propagated by the previous sub-modules. We check our model with three state-of-the-art (SOTA) datasets in SISR and compare the results with the SOTA models. The results show that our model is the only one able to manage our wider set of deformations. Furthermore, our model overcomes all current SOTA methods for a standard set of deformations. In terms of computational load, our model also improves on the two closest competitors in terms of efficiency. Although the approach is non-blind and requires an estimation of the blur kernel, it shows robustness to blur kernel estimation errors, making it a good alternative to blind models.