I'm trying to perform image recognition via a SIFT algorithm using solid shapes.
To improve the performance, I would like to take the image (captured from an iOS camera), make all non-light colors a dark red and make the remaining pixels transparent.
An example of a before and after of what I'm trying to achieve is attached in the images on this question.
Assumption: the images are always solid-black shapes and printed on plain white paper
Can someone please help me with this or point me in the right direction?