0

I have an OpenCV application and I have to implement a correspondence search using varying support weights between two image pair. This work is very similar to "Adaptive Support-Weight Approach for Correspondence Search" by Kuk-Jin Yoon and In So Kweon. The support weights are in a given support window. I calculate dissimilarity between pixels using the support weights in the two images. Dissimilarity between pixel 'p' and 'Pd' is given by Fig.1 - dissimilarity formula
where Pd and Qd are the pixels in the target image when pixels p and q in the reference image have a disparity value d; Np and Npd are the support weight.

After this, the disparity of each pixel is selected by the WTA (Winner-Takes-All) method as: Fig.2 - WTA method

What I would like to know is how to proceed starting with the formula of the fig.1 (function computing dissimilarity and weights that I have written), i.e. which pixel to consider? Where to start? What pixel with? Any suggestion? The final result of the work should be similar to: Fig.3 - Tsukuba
What could be a good way to do it?

UPDATE1
Should I start creating a new image, and then consider the cost between the pixel (0,0) and all the other pixels, find the minimum value and set this value as the value in the new image at pixel (0,0) ? And so on with the other pixels?

Franktrt
  • 373
  • 1
  • 8
  • 18
  • We like to see some effort: [MCVE](http://stackoverflow.com/help/mcve). What have you tried? – Berriel Mar 11 '16 at 00:19
  • @berriel I don't ask code, but I'd like to know a theoretical approach or suggestion about how to go on, cause I already computed the dissimilarity between two pixel. Moreover this paper is part of a bigger work. :) – Franktrt Mar 11 '16 at 00:26
  • Signal processing knowledge-based (i.e. with no code) are off-topic here. Try asking on DSP StackExchange instead: http://dsp.stackexchange.com. FWIW, I actually don't understand your question. The simplest way to calculate disparities is to aggregate a sum of squared differences between a patch in the source image with a patch in the target image that is horizontally displaced by a certain disparity. You do this for a set of disparities and choose the minimum SSD. In their work, it is now a **weighted** sum. Simply change the SSD so that each pixel is weighted by some weighting function. – rayryeng Mar 11 '16 at 09:50
  • @rayryeng Ok. Let's consider I find the minimimum cost for pixel (x, y) for the entire row x. Do I set this value as the same position pixel in the new gray image? – Franktrt Mar 11 '16 at 10:50

0 Answers0