System and Methods for Image Restoration
Tufts University
posted on 09/02/2011
Tufts University Categories:
imaging
software
Additional Keywords:
(none specified)
Invention Summary
Background: Great libraries of historic texts, manuscripts, and images are available at this time. The information contained in these documents will be of critical importance in documenting the history of humanity in the years to come. However, many of these documents are degrading for a variety of reasons, including semi-transparent blotches. Current restoration techniques are either chemical in nature or involve significant user interaction through a desktop computer. Due to operator costs, either method is too expensive for the large quantity of information that needs to be preserved. Automated blotch removal algorithms exist, however they all have severe limitations: Some mechanisms only work on text and not images while others leave edges around the blotch or blur the image within the blotch. Invention: This invention introduces an automated Local Logarithmic Restoration Algorithm to restore semi-transparent blotches in a variety of documents. Intensive computer simulations show: (1) the new algorithm outperforms competing algorithms by providing better local contrast preservation, expanded saturation reduction capabilities, and improved edge removal; and (2) it is a flexible and practical algorithm for batch restoration many pages of text, manuscripts, or images.
File Number: T001533
Other Information:
Investigator(s)
Karen A. Panetta
Contact
Martin Son
617-636-3605
martin.son@tufts.edu, Martin Son
617-636-3605
martin.son@tufts.edu
This innovation currently is not available for online licensing. Please contact the case manager at Tufts University for more information.
Find more innovations
