摘要 |
<p>An algorithm is provided for enhancing the resolution of various types of data such as image data. The data being enhanced can be data which has been sampled using a locally inhomogeneous sensitivity pattern such as a Bayer-type color mosaic, an exposure-level mosaic, or a combined color and exposure-level mosaic. The resolution enhancement algorithm uses high resolution and low resolution versions of a sample image to 'learn' (i.e., optimize) a set of parameters of a mapping from the low resolution data to the high resolution data. The mapping can comprise a polynomial model. The learning procedure is repeated for a variety of sample images. The learned mapping can then be used to process other low resolution images to obtain corresponding, higher resolution images. The learned mapping provides resolution enhancement which is superior to that of conventional interpolation techniques.</p> |