Lowlights map

Correct image background using 3d visibility operators.
63 Downloads
Updated 28 Apr 2019

View License

Implementation of the algorithm of : Kligler, N., Katz, S., & Tal, A. (2018). Document Enhancement Using Visibility Detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 2374-2382).
This algorithm use the 3d visibility operator : Hidden point removal (HPR) or Target Point Occluding (TPO) to find deep valley points (black caracters) in a grayscale image.
The miniature is generated by :

I = imread('PR3.png');
Ilab = rgb2lab(I);
Img = mat2gray(Ilab(:, :, 1));
L = LowLight(Img, [5e-2, 1], 1e-4);
figure;
imshowpair(Img, L, 'montage')
title('Input and corrected image')

The 'PR3.png' image come from de DIBCO2011 database and is found here : http://utopia.duth.gr/~ipratika/DIBCO2011/benchmark/

Cite As

Geoffrey Schivre (2026). Lowlights map (https://nl.mathworks.com/matlabcentral/fileexchange/71388-lowlights-map), MATLAB Central File Exchange. Retrieved .

MATLAB Release Compatibility
Created with R2018a
Compatible with R2013a and later releases
Platform Compatibility
Windows macOS Linux
Version Published Release Notes
1.0.1

Correct the description of L in LowLight function.

1.0.0