Image Reconstruction and Regeneration (iInp)

Imatrics Inpainting & Super-resolution (iInp) can regenerate impaired areas of a picture or eliminate undesired components of a digital image:

• Dead pixels (saturation on acquisition, losses in transmission).
• Scratches or stains on old movies.
• Missing parts of pictures on other media (paintings, old photographs).
• Streetlights, cables and traffic signs in landscapes.
• Parts of an image containing information that should not be revealed.

iInp can also improve the quality of low-resolution images enlarged using digital zoom or integrated with higher-resolution images. It can also reduce the visual impact of quantification artefacts (e.g. from using a colour pallet in formats such as GIF or BMP).

Result of applying iInp to an image which had lost 40% of its pixels (image from [1])

Applying iInp to a picture degraded due to the use of a colour palette (GIF, a common image format on the Internet)

Example of applying iInp to an image to improve resolution (image from [1])

[1] L. Mancera, ”Sparse Approximation to Natural Images based on non-Convex Optimisation and Aplications to Restoration”, Ph.D. DIssertation, Univ. de Granada, March 2008.