Each pixel corresponds to a residue in the horizontal sequence, and another residue in the vertical sequence. The pixel's color depends on how similar the two sequences are around these two positions: each pixel represents a score, with high scores meaning good matches (in this case, the darker the pixel, the lower the score). Obviously there will be a large number of pixels with low scores and only a few ones with high scores. It is likely that you will have to tune the grayscale in order to make the background noise (low scores) disappear and the similar regions stand out more clearly (the image shown in Fig. 3 has quite a bit of background). To do this, use the histogram window (Fig. 4):