PointNet分类和分割代码复现（跑程序）
分类
代码地址以及相关操作：(https://github.com/charlesq34/pointnet)
分类比较简单，按照网上的步骤很少会出问题，github上边也比较详细，就不具体说明了。
分割
我在运行分割时遇到两个问题
第一：找不到文件夹
FileNotFoundError: [Errno 2] No such file or directory: '/home/l/Desktop/pointnet-master/part_seg/./PartAnnotation/03001627/points/355fa0f35b61fdd7aa74a6b5ee13e775.pts'
这个因为数据集下载的时候出错了，一共两个数据集，如果运行 sh download_data.sh
下载不下来，可以通过链接下载好，在去运行程序，
train.py运行效果如下：
<<< Testing on the test dataset ...
Loading test file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_val0.h5
Testing Total Mean_loss: 14097378435.845493
Testing Label Mean_loss: 528065397.047210
Testing Label Accuracy: 0.211373
Testing Seg Mean_loss: 14097378435.845493
Testing Seg Accuracy: 0.016106
Category Airplane Object Number: 389
Category Airplane Label Accuracy: 1.000000
Category Airplane Seg Accuracy: 0.000000
Category Bag Object Number: 8
Category Bag Label Accuracy: 0.000000
Category Bag Seg Accuracy: 0.000000
Category Cap Object Number: 5
Category Cap Label Accuracy: 0.000000
Category Cap Seg Accuracy: 0.000000
Category Car Object Number: 79
Category Car Label Accuracy: 0.000000
Category Car Seg Accuracy: 0.000000
Category Chair Object Number: 395
Category Chair Label Accuracy: 0.000000
Category Chair Seg Accuracy: 0.000000
Category Earphone Object Number: 6
Category Earphone Label Accuracy: 0.000000
Category Earphone Seg Accuracy: 0.000000
Category Guitar Object Number: 78
Category Guitar Label Accuracy: 0.000000
Category Guitar Seg Accuracy: 0.000000
Category Knife Object Number: 35
Category Knife Label Accuracy: 0.000000
Category Knife Seg Accuracy: 0.000000
Category Lamp Object Number: 142
Category Lamp Label Accuracy: 0.000000
Category Lamp Seg Accuracy: 0.211065
Category Laptop Object Number: 44
Category Laptop Label Accuracy: 0.000000
Category Laptop Seg Accuracy: 0.000000
Category Motorbike Object Number: 26
Category Motorbike Label Accuracy: 0.000000
Category Motorbike Seg Accuracy: 0.000000
Category Mug Object Number: 16
Category Mug Label Accuracy: 0.000000
Category Mug Seg Accuracy: 0.000000
Category Pistol Object Number: 30
Category Pistol Label Accuracy: 0.166667
Category Pistol Seg Accuracy: 0.001693
Category Rocket Object Number: 8
Category Rocket Label Accuracy: 0.000000
Category Rocket Seg Accuracy: 0.000000
Category Skateboard Object Number: 15
Category Skateboard Label Accuracy: 0.000000
Category Skateboard Seg Accuracy: 0.000000
Category Table Object Number: 588
Category Table Label Accuracy: 0.000000
Category Table Seg Accuracy: 0.000000
>>> Training for the epoch 0/200 ...
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train1.h5
Training Total Mean_loss: 2.208344
Training Label Mean_loss: 3.707007
Training Label Accuracy: 0.030273
Training Seg Mean_loss: 1.779380
Training Seg Accuracy: 0.546435
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train4.h5
Training Total Mean_loss: 1.569824
Training Label Mean_loss: 3.564084
Training Label Accuracy: 0.027832
Training Seg Mean_loss: 1.324413
Training Seg Accuracy: 0.633967
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train0.h5
Training Total Mean_loss: 0.963046
Training Label Mean_loss: 3.551116
Training Label Accuracy: 0.027344
Training Seg Mean_loss: 0.925087
Training Seg Accuracy: 0.726815
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train2.h5
Training Total Mean_loss: 0.797747
Training Label Mean_loss: 3.585615
Training Label Accuracy: 0.032715
Training Seg Mean_loss: 0.775099
Training Seg Accuracy: 0.766800
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train5.h5
Training Total Mean_loss: 2.528071
Training Label Mean_loss: 3.569277
Training Label Accuracy: 0.026899
Training Seg Mean_loss: 0.916843
Training Seg Accuracy: 0.723155
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train3.h5
Training Total Mean_loss: 3.322942
Training Label Mean_loss: 3.470755
Training Label Accuracy: 0.033203
Training Seg Mean_loss: 1.223184
Training Seg Accuracy: 0.650061
<<< Testing on the test dataset ...
Loading test file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_val0.h5
Testing Total Mean_loss: 1.655914
Testing Label Mean_loss: 6.365255
Testing Label Accuracy: 0.019850
Testing Seg Mean_loss: 1.061547
Testing Seg Accuracy: 0.743527
Category Airplane Object Number: 389
Category Airplane Label Accuracy: 0.025707
Category Airplane Seg Accuracy: 0.761382
Category Bag Object Number: 8
Category Bag Label Accuracy: 0.000000
Category Bag Seg Accuracy: 0.000000
Category Cap Object Number: 5
Category Cap Label Accuracy: 0.000000
Category Cap Seg Accuracy: 0.003223
Category Car Object Number: 79
Category Car Label Accuracy: 0.000000
Category Car Seg Accuracy: 0.755044
Category Chair Object Number: 395
Category Chair Label Accuracy: 0.000000
Category Chair Seg Accuracy: 0.595826
Category Earphone Object Number: 6
Category Earphone Label Accuracy: 0.000000
Category Earphone Seg Accuracy: 0.000000
Category Guitar Object Number: 78
Category Guitar Label Accuracy: 0.025641
Category Guitar Seg Accuracy: 0.855669
Category Knife Object Number: 35
Category Knife Label Accuracy: 0.000000
Category Knife Seg Accuracy: 0.029408
Category Lamp Object Number: 142
Category Lamp Label Accuracy: 0.147887
Category Lamp Seg Accuracy: 0.644246
Category Laptop Object Number: 44
Category Laptop Label Accuracy: 0.000000
Category Laptop Seg Accuracy: 0.521717
Category Motorbike Object Number: 26
Category Motorbike Label Accuracy: 0.000000
Category Motorbike Seg Accuracy: 0.642597
Category Mug Object Number: 16
Category Mug Label Accuracy: 0.000000
Category Mug Seg Accuracy: 0.722504
Category Pistol Object Number: 30
Category Pistol Label Accuracy: 0.133333
Category Pistol Seg Accuracy: 0.892220
Category Rocket Object Number: 8
Category Rocket Label Accuracy: 0.000000
Category Rocket Seg Accuracy: 0.509094
Category Skateboard Object Number: 15
Category Skateboard Label Accuracy: 0.000000
Category Skateboard Seg Accuracy: 0.675228
Category Table Object Number: 588
Category Table Label Accuracy: 0.000000
Category Table Seg Accuracy: 0.923971
>>> Training for the epoch 1/200 ...
Loading train file /home/l/Desktop/pointnet-master/part_seg/./hdf5_data/ply_data_train3.h5
Training Total Mean_loss: 1.112694
Training Label Mean_loss: 3.481067
Training Label Accuracy: 0.041992
Training Seg Mean_loss: 0.951263
Training Seg Accuracy: 0.713590
第二：在运行过程中会显示：资源消耗完了（翻译过来大概）
这个是batch_size太大了，去程序里面改小，8或者4,就可以啦。
test.py运行结果如下：
Loading model train_results/trained_models/epoch_190.ckpt
Model restored.
0/2874 ...
100/2874 ...
200/2874 ...
300/2874 ...
400/2874 ...
500/2874 ...
600/2874 ...
700/2874 ...
800/2874 ...
900/2874 ...
1000/2874 ...
1100/2874 ...
1200/2874 ...
1300/2874 ...
1400/2874 ...
1500/2874 ...
1600/2874 ...
1700/2874 ...
1800/2874 ...
1900/2874 ...
2000/2874 ...
2100/2874 ...
2200/2874 ...
2300/2874 ...
2400/2874 ...
2500/2874 ...
2600/2874 ...
2700/2874 ...
2800/2874 ...
Accuracy: 0.926577
IoU: 0.816931
02691156 Total Number: 341
02691156 Accuracy: 0.9127049599924395
02691156 IoU: 0.8307283910488453
02773838 Total Number: 14
02773838 Accuracy: 0.948321955544608
02773838 IoU: 0.7573385238647461
02954340 Total Number: 11
02954340 Accuracy: 0.8903708891435103
02954340 IoU: 0.7879825938831676
02958343 Total Number: 158
02958343 Accuracy: 0.9067590447920787
02958343 IoU: 0.7461091053636768
03001627 Total Number: 704
03001627 Accuracy: 0.9405029470270331
03001627 IoU: 0.8934942592274059
03261776 Total Number: 14
03261776 Accuracy: 0.9187915665762765
03261776 IoU: 0.7400967734200614
03467517 Total Number: 159
03467517 Accuracy: 0.9655065956355641
03467517 IoU: 0.9105948202265134
03624134 Total Number: 80
03624134 Accuracy: 0.9240625381469727
03624134 IoU: 0.8528668403625488
03636649 Total Number: 286
03636649 Accuracy: 0.900685450413844
03636649 IoU: 0.7965391465833971
03642806 Total Number: 83
03642806 Accuracy: 0.977536672569183
03642806 IoU: 0.9512975302087255
03790512 Total Number: 51
03790512 Accuracy: 0.8602656196145451
03790512 IoU: 0.6449186287674249
03797390 Total Number: 38
03797390 Accuracy: 0.9925226914255243
03797390 IoU: 0.9128187079178659
03948459 Total Number: 44
03948459 Accuracy: 0.9535756544633345
03948459 IoU: 0.8133323842828925
04099429 Total Number: 12
04099429 Accuracy: 0.8038069407145182
04099429 IoU: 0.5588903824488322
04225987 Total Number: 31
04225987 Accuracy: 0.9447346964190083
04225987 IoU: 0.7312387651012789
04379243 Total Number: 848
04379243 Accuracy: 0.9219124991938753
04379243 IoU: 0.7494366483868293