There is low Peak-to-Sidelobe Ratio in the method

[zeroRains]

I have read the paper and used the calculate method of PSR to measure the code of this repository. I found that the psr is lower than the paper say. MOSSE usually has 20 ~60 psr. But this code only has 7 psr in your demo. Why is this?

[jsnewton3]

I have read the paper and used the calculate method of PSR to measure the code of this repository. I found that the psr is lower than the paper say. MOSSE usually has 20 ~60 psr. But this code only has 7 psr in your demo. Why is this?

Hey! Out of curiosity, how did you calculate the PSR? Getting the max value is easy, but how do you get the mean and sigma of the "sidelobe"? In the paper, it says that the sidelobe "is the rest of the pixels excluding an11×11 window around the peak".