Not use reasonable set for training?

[Jokoe66]

@liuwei16 Hi, thanks for your research. In generate_cache_city.py, you cache boxes with label=1 and height >= 50, but do not consider the occlusion level, that means you are not using the reasonable training set. Is it right? If so, what do you think of this setting?

[frezaeix]

Hi, Thanks for sharing your code. I have also a similar question. It seems that in your cache file for training on CityPersons, you did not consider occlusion. So, did you report your results on the paper based on this training set?

[bomtorazek]

For Caltech, the provided cache files for training from authors are also different from those from running the given code. From given, it says num of ped and emp samples: 9105 33677 From generated from the code, it says num of ped and emp samples: 22356 20426

I believe that authors omitted some lines regarding occlusion levels by mistake

[Boisa]

细思极恐，作者会不会在测试集上也忽略了一些遮挡的样本？