Please use this identifier to cite or link to this item:
https://open.uns.ac.rs/handle/123456789/6659
Title: | Microscopy image enhancement for cost-effective cervical cancer screening | Authors: | Lindblad J. Bengtsson E. Sladoje N. |
Issue Date: | 1-Jan-2015 | Journal: | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Abstract: | © Springer International Publishing Switzerland 2015. We propose a simple and fast method for microscopy image enhancement and quantitatively evaluate its performance on a database containing cell images obtained from microscope setups of several levels of quality. The method utilizes an efficiently and accurately estimated relative modulation transfer function to generate images of higher quality, starting from those of lower quality, by filtering in the Fourier domain. We evaluate the method visually and based on correlation coefficient and normalized mutual information. We conclude that enhanced images exhibit high similarity, both visually and in terms of information content, with acquired high quality images. This is an important result for the development of a cost-effective screening system for cervical cancer. | URI: | https://open.uns.ac.rs/handle/123456789/6659 | ISBN: | 9783319196640 | ISSN: | 03029743 | DOI: | 10.1007/978-3-319-19665-7_37 |
Appears in Collections: | FTN Publikacije/Publications |
Show full item record
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.