A similarity matrix is very similar to a correlation matrix. A similarity matrix determines the how close the two data points are. In the picture above the pixels of 512 photographs was analyzed. Photographer did this by grouping pictures from similar events into distinct categories and then running a similarity matrix to see how the pixels of the different pictures corresponded. The blue pixels show a correlation of 0 while the red pixels show a correlation of 1.
No comments:
Post a Comment