Wisely utilizing the internal and external learning methods is a newchallenge in super-resolution problem. To address this issue, we analyze theattributes of two methodologies and find two observations of their recovereddetails: 1) they are complementary in both feature space and image plane, 2)they distribute sparsely in the spatial space. These inspire us to propose alow-rank solution which effectively integrates two learning methods and thenachieves a superior result. To fit this solution, the internal learning methodand the external learning method are tailored to produce multiple preliminaryresults. Our theoretical analysis and experiment prove that the proposedlow-rank solution does not require massive inputs to guarantee the performance,and thereby simplifying the design of two learning methods for the solution.Intensive experiments show the proposed solution improves the single learningmethod in both qualitative and quantitative assessments. Surprisingly, it showsmore superior capability on noisy images and outperforms state-of-the-artmethods.
展开▼
