the false positive rate & detection rate setting for each classification in the cascade framework

luoyetx / JDA

C++ implementation of Joint Cascade Face Detection and Alignment.

BSD 3-Clause "New" or "Revised" License

184 stars 140 forks source link

the false positive rate & detection rate setting for each classification in the cascade framework #27

Open hesperxxz opened 7 years ago

hesperxxz commented 7 years ago

Hi luoyetx, thanks for your code. In your config.json, for each cart, detection rate = 1 - drop_n / number_of_positive_samples = 1-2/50000 = 0.99996 max false positive rate = 1 - restart.th = 0.999 Therefore, the final detection rate is pow(0.99996, 5405) = 0.8976 the final false positive rate = pow(0.999, 5405) = 0.0671 Is that right? Maybe the DR is a little low and the FPR is too high.

luoyetx commented 7 years ago

@hesperxxz your formula is right. And I think I'm using number_of_positive_samples = 100,000 (not all of them has landmark points). And I only trained 2700 carts.

The FPR you calculates is not the truth. Actually max false positive rate < 1 - restart.th, it will be much smaller than 0.999. I think the final FR will be 1e-5, that small.