Recognizing numbers from the low-resolution patterns in digital images

In this paper, we suggest a new method for solving a problem of recognition numbers from low-resolution (LR) patterns in digital images. Having the digital photographs, for example, with the cars, we can recognize the numbers in the license plates of the cars, even if the plate patterns are still unclear after applying the Super-Resolution. The proposed method is based on statistical analysis and learning the features of the single LR digit pattern with the features of high-resolution (HR) one.