Hello, why do I get so bad result?

[shoutOutYangJie]

I test one image, but result is so bad. There are so many false-positive samples. Why?

[kakusikun]

how about to use a probability threshold to filter the result? :p (this is p=0.5 result)

[shoutOutYangJie]

how about to use a probability threshold to filter the result? :p (this is p=0.5 result)

wawa,so great! so i need to set score from 0.1 to 0.5? right?

[kakusikun]

It seems that the score in bbox_process is the threshold to filter the output of seman_map which might be the probability of object in pixel level. I'm not sure the final probability of bbox is the probability of seman_map, you can try it. ^^

[Boisa]

我测试了一张图像，但是结果太差了。假阳性样本太多。为什么？

Hello!How did you test a single picture？

I test one image, but result is so bad. There are so many false-positive samples. Why?

Hello!How did you test a single picture？How should the code be modified，Or did you write the script by yourself?

[SouleTraveler]

我测试了一张图像，但是结果太差了。假阳性样本太多。为什么？

Hello!How did you test a single picture？

I test one image, but result is so bad. There are so many false-positive samples. Why?

Hello!How did you test a single picture？How should the code be modified，Or did you write the script by yourself? 你好，我有win10运行test_caltech.py时，报错 ModuleNotFoundError: No module named 'nms' 您知道这个问题怎样解决吗？不胜感激