| 主权项 |
1. A method comprising:
receiving, by a processing device, an input image patch of an input image; generating, by the processing device, a first interim output image patch by applying, to the input image patch, a first enhancement operation using self-examples; generating, by the processing device, a second interim output image patch by applying, to the input image patch, a second enhancement operation using external examples; determining, by the processing device, a first weight and a second weight based on (i) a content metric for the input image patch that indicates an amount of smoothness or texture in an up-sampled input image patch that is included in an up-sampled version of the input image and that corresponds to the input image patch, (ii) a first distance between a first vector representing the input image patch and a second vector representing a self-example image patch obtained from a self-example image, and (iii) a second distance between the first vector representing the input image patch and a third vector representing an external-example image patch obtained from an external example image; and generating, by the processing device, an output image patch by combining the first interim output image patch modified using the first weight and the second interim output image patch modified using the second weight. |
