Help Starting Project – Mapping Pictures to GL World

Dec 8, 2013 at 6:37am

Help Starting Project – Mapping Pictures to GL World

Hi everyone – I’ve programmed quite a bit of Max but am a little at a loss for the initiatory steps for an idea I have, as my jitter skills are somewhat limited. I’d like to create a patch that can take a picture, analyze it, and map GL lines to recreate the lines in the picture. I’d imagine I’d begin by using jit.rgb2luma, and then analyze the difference between pixels and at points where pixels have a higher difference, you could map a line. I know there has to be some jitter math whereby I could analyze a matrix and determine these locations in a somewhat succinct way. I’m not an “expr”ert, so I’m not quite sure of the math. Could someone push me in the right direction? Thanks!

#275107
Dec 8, 2013 at 10:44pm

“cv.jit is a collection of max/msp/jitter tools for computer vision applications.”

cv.jit is designed for webcam/video but I am almost certain that it can be repurposed for still images.

“A patch that can take a picture, analyze it, and map GL lines to recreate the lines in the picture.”

I did some digging, and I discovered that the cv.jit.snake object does this entire concept quite well!

“A snake is a sequence of points, or anchors, that seeks to fit itself to edges in an image while trying to maintain its original shape. This achieves something similar to “magnetic lasso” or “intellect scissors” of some image-editing software. It is defined by a one dimension, 5plane matrix.”

http://jmpelletier.com/cvjit/

#275169

You must be logged in to reply to this topic.