Extremely slow imshow() when using 16-bit images
Afficher commentaires plus anciens
I'm starting by taking a matrix of data from a DICOM file and making it indexed (so that I can apply a user-specified colormap).
shades = 2^bit_val
grayData = mat2gray(dicomData);
% Limiting to shades-1 to allow for a special mark-up color
indexedData = gray2ind(data, shades-1);
%Then code to calculate cmap with the special color at the end
...
imshow(indexedData, cmap, handles.Axes);
If bit_val is 8, then the code takes 0.01 seconds to display which is fine. However, if bit_val is 16, then it takes nearly a whole second to display.
The primary issue is that I want to be able to quickly/smoothly switch between different slices being displayed.
Is there any way to improve the rendering time?
Edit:
According to Wikipedia's entry for palette ( http://en.wikipedia.org/wiki/Palette_%28computing%29#System_and_logical_palettes_under_Microsoft_Windows )
Under Windows, "When a given application intends to output colorized graphics and/or images, it can set their own logical palette, that is, its own private selection of colors (up to 256)."
Perhaps this explains why an 8-bit color map (with only 256 colors) displays faster than a 16-bit color map (with 65536 colors) on Windows.
3 commentaires
Walter Roberson
le 12 Déc 2012
Which platform are you using? I have read that MS Windows is not designed to be able to handle colormaps with more than 256 entries.
Sean de Wolski
le 12 Déc 2012
How are you timing it?
Christopher
le 14 Déc 2012
Réponses (1)
Sean de Wolski
le 12 Déc 2012
Modifié(e) : Sean de Wolski
le 12 Déc 2012
X = uint16(rand(1000)*60000);
t = 0;
figure;ax = axes;
for ii = 1:10
tic;
imshow(X,'parent',ax);
drawnow;
t = t+toc;
cla;
drawnow;
end
disp(t/10);
% 0.0335 on my system
Also Have you seen implay()?
doc implay
4 commentaires
Christopher
le 14 Déc 2012
Modifié(e) : Christopher
le 14 Déc 2012
Image Analyst
le 14 Déc 2012
You don't need to put tic and toc inside a disp() function.
Sean de Wolski
le 17 Déc 2012
@Christopher: Put this in a loop (like I did) and run it in a script file. That way you don't get any overhead that may be introduced by the command window.
Christopher
le 17 Déc 2012
Modifié(e) : Christopher
le 17 Déc 2012
Catégories
En savoir plus sur Red dans Centre d'aide et File Exchange
Produits
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!