keineahnung2345
keineahnung2345
Yes, I hope this PR could be reviewed ASAP. @waleedka
@pmaksud It should work fine with bg image input. BG images only contribute to rpn_class_loss and nothing else.
Since MaskRCNN is composed of many different parts, so it's hard to say which part of the entire model goes wrong. There is another issue discussing the interaction of RPN,...
You can check: [彙整中文與英文的詞性標註代號：結巴斷詞器與FastTag / Identify the Part of Speech in Chinese and English](http://blog.pulipuli.info/2017/11/fasttag-identify-part-of-speech-in.html).
It also requires the package "mock", so one should "pip install mock dill" before running this.
Have added a test that extract all.pcd, the effect is like: 
@mvieth I've noticed that formatting check would not be done on `filters` module, should it be enabled?
I see, that's not an easy work.
Try this: https://github.com/sshaoshuai/PointRCNN/pull/184
@mvieth Any updates?