from Konstantinos Zagoris
In the frame of “READ” project, document image enhancement tools have been implemented to be used as a preprocessing stage for the binarisation algorithm. After experimental work, the enhancement tools consists of two different modules.
The first module is an anisotropic diffusion process or also called Perona–Malik diffusion [Perona1987]. It is a procedure aiming at reducing image noise without removing significant parts of the image content - typically edges.
The second module is inspired by the characteristics of the ganglion cells of the Human Visual System (HVS) [Nelson2004, Vonikakis2011]. It can deal with various types of degradations, such as uneven illumination, shadows, low contrast, smears, and heavy noise densities.
The document image enhancement tool is developed in C++11 and it is available at github under LGPL-3.0
| Original Document | Binarization |
|---|---|
 |
 |
| Enhancement | Binarisation after enhancement |
|---|---|
 |
 |
Usage: enhancement [infile] [outfile]
[Nelson2004] Nelson R, Kolb H. 2004. ON and OFF pathways in the vertebrate retina and visual system. In: Chalupa LM, Werner JS (eds) The visual neurosciences. MIT Press, Cambridge, pp 260–278
[Perona1987] Pietro Perona and Jitendra Malik. 1987. "Scale-space and edge detection using anisotropic diffusion". Proceedings of IEEE Computer Society Workshop on Com-puter Vision,. pp. 16–22.
[Vonikakis2011] Vonikakis, V., Andreadis, I., & Papamarkos, N. 2011. Robust document bi-narisation with OFF center-surround cells. Pattern Analysis and Applications, 14(3), 219-234.