When applying threshold on a picture, it's often difficult to set it correctly to keep most details (so that darker/lighter parts of the image don't get entirely black/white).<br/>
I experimented a bit with the idea of an optimized threshold.<br/>
The best approach so far was to compare each pixel to an average value of the neighboring pixels (within a radius of around 200 pixels), rather than to a fixed value.<br/>
I used only a few pixels (just for performance, because it isn't gpu-based yet) and applied a different ratio based on the distance.<br/>
Then, just for fun I tried with one threshold on each RGB layer to get 8 colors pics.<br/>
I suppose this is in some way a cheap version of some kernel-based filters or dithering techniques, and there is probably a name for it that I don't know about.<br/>
<br/>
<img src ="Hieronymus_Bosch_051.jpg" style="width:600px;" /><br/>
<br/>
Here is a standart threshold applied to a picture :<br/>
<img src ="resultPh1.jpg" style="width:600px;" /><br/>
And now an "optimized" threshold with various settings :<br/>
<img src ="result4.png" style="width:600px;" /><br/>
The color of the dress seems to change dependin on the threshold bias.<br/>
This one with a very narrow range :<br/>
<img src ="result3.png" style="width:600px;" /><br/>
One with colors :<br/>
<img src ="result5.png" style="width:600px;" /><br/>
<br/>
<a href="optimThreshold.zip">the source</a><br/>
<br/>
My appoligies to you, the woman on the picture, I don't know who you are and just picked a random photo to make tests on it.<br/>
<br/>