Comparison of Signatures

Next, we can compare these two signatures, to see whether they have mapped our different images to different signature values.

You can compare the signatures with one simple line of R code as follows:

comparison<-mean(abs(borges_signature-building_signature))

This comparison takes the absolute value of the difference between each element of the two signatures, and then calculates the mean of those values. If two signatures are identical, then this difference will be 0. The larger the value of comparison, the more different the two images are.

In this case, the value of comparison is 0.644, indicating that on average, the corresponding signature entries are about 0.644 apart. This difference is substantial for a dataset where the values only range between 1 and -1. So we see that our method for creating signatures has created very different signatures for very different images, as we would expect.

Now, we can calculate a signature for an image that is very similar...