Super-resolution is a technology of generating high-resolution images from multiple low-resolution images through image processing. Different from the traditional spatial scaling techniques (such as bi-cubic interpolation and poly-phase interpolation), super-resolution utilizes both spatial and temporal information from multiple low-resolution images. Super-resolution is capable of producing images with much more spatial details and sharper edges, and with fewer artifacts (such as blurriness and aliasing) and less noticeable noise.

Shown in the following are some examples produced by the super-resolution algorithm developed by us. Note that all the results shown below were achieved without any assumptions as to what pre-processing had been applied to the input images.

super-resolution examples

super-resolution examples (with sharpening)