CANDID Conclusions
Captures information in a manner similar to histograms, but no need for discrete bins
Not viable for large problems -- does not scale well
Probability Density Function are SLOW to compute
Subtracting a background improves system performance when inner product