fanhqme / pointsetgeneration Goto Github PK

Hello, I read some paper for face 3d reconstruction, but they are based on 3DMM. Do you think your model can generate 3d face??? Is it better than 3DMM???

I can't access the download webpage in google drive. just appear the several links with access Denied.

Anyone has the answer.

I hope you can give me access. Thank you very much.

How to generate the corresponding mask ? Such as 1_m.png .

I am using python3 and tensorflow1.1.2 and it can not work.

When I run the code python train__nn.py, I get the following error:
File "train_nn.py", line 228
print bno,'t',trainloss_accs[0]/trainloss_acc0,trainloss_accs[1]/trainloss_acc0,trainloss_accs[2]/trainloss_acc0,'v',validloss_accs[0]alidloss_acc0,validloss_accs[1]alidloss_acc0,validloss_accs[2]alidloss_acc0,total_loss-showloss,t1-t0,t2-t1,time.time()-t0,fetchworker.queue.qsize()
^

How do I solve with it?

The three mentioned in the article‘ .PKL ’file ,I do not have permission to access, hope to get the author's approval or permission

Hi,
I tried the code, most things went fine, however, there was an error showing that data/0/64.gz didn't exist. Could you please provide the link to download complete training data or any solutions to the issue? Thanks!

The following is the error message.

error! data file not exists: ../data/0/64.gz
please KILL THIS PROGRAM otherwise it will bear undefined behaviors
Exception in thread Thread-1:
Traceback (most recent call last):
File "/usr/lib/python2.7/threading.py", line 801, in __bootstrap_inner
self.run()
File "/home/twke/repos/PointSetGeneration/depthestimate/BatchFetcher.py", line 70, in run
self.queue.put(self.work(self.bno%300000))
File "/home/twke/repos/PointSetGeneration/depthestimate/BatchFetcher.py", line 37, in work
assert False,"data file not exists: %s"%path
AssertionError: data file not exists: ../data/0/64.gz

(fastai-3.7) D:\git_hub\3D\PointSetGeneration\demo>python view.py r1.png.txt
Traceback (most recent call last):
File "view.py", line 4, in
a=np.loadtxt(sys.argv[1])
File "C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\numpy\lib\npyio.py", line 1101, in loadtxt
for x in read_data(_loadtxt_chunksize):
File "C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\numpy\lib\npyio.py", line 1028, in read_data
items = [conv(val) for (conv, val) in zip(converters, vals)]
File "C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\numpy\lib\npyio.py", line 1028, in
items = [conv(val) for (conv, val) in zip(converters, vals)]
File "C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\numpy\lib\npyio.py", line 746, in floatconv
return float(x)
ValueError: could not convert string to float: '-0.07616415-0.1313511'

I have some questions.
1.How to train a new network(use the data provided by author)
Must I download the data from [https://shapenet.cs.stanford.edu/media/sampledata_220k.tar] ?
Can I try less data,such as 64.gz they provided? Seems like the detail is in the BatchFectcher.py. Is there some methods that I can modify the BatchFectcher.py?
2.Have you try to train the network with your own data? And how to prepare it?
I would appreciate it if you could give me some information.

It seems that you are using nearest neighbour distance to calculate loss instead of earth mover's distance in your code, which is inconsistent with your paper section 4.3. Am I missing something here?

Thanks,

Lei

Do you know how to replace the training data？ And how to extract the .gz file? Thanks!

Hello, great work!
Where is MoN and random input?

Hi, Thank you for your work, I just want to kown, how can I display the predicted point cloud in batches with the provided trained model?
Wish to your reply！

Great work! Thanks you for posting the code!

When I use your code and compile cuda code. There is a message shows: make: Circular depthestimate/tf_nndistance_g.cu <- depthestimate/tf_nndistance_g.cu.o dependency dropped.

I am not sure if this is an error message. But when I try to train the network an error shows up: NotFoundError: ./tf_nndistance_so.so: undefined symbol: _ZTIN10tensorflow8OpKernelE

I think this may be because I didn't successfully build the code.

Is there anyone run into the same problem?

Also could the author please post the version of tensorflow, g++, tflearn etc.? In case there are version issues.

Thank you very much!

In the ImageCaptureGuide.pdf, the mask image should be prepared first. But I don't see the method to generate it. Could you or anyone help me?
Thanks a lot!

Hi

While trying to compile the code (as provided), following error occurs

depthestimate/tf_nndistance_g.cu:4:61: fatal error: third_party/eigen3/unsupported/Eigen/CXX11/Tensor: No such file or directory

The system on which the code is being compiled has tensorflow, cuda 8, cudnn installed and working properly. The required package is also available at

/home/user/tensorflow/lib/python2.7/site-packages/tensorflow/include/third_party/eigen3/unsupported/Eigen/CXX11/

Could you kindly inform if there are any specific settings that need to be done for the code to find appropriate files ?

I compiled the makefile succesfully

When I run python train_nn.py I get this error:
tensorflow.python.framework.errors_impl.NotFoundError: ./tf_nndistance_so.so: undefined symbol: _ZN10tensorflow12OpDefBuilder5InputESs

I am using a conda environment with python2.7, tensorflow 1.4 and gcc 4.8

This is my makefile:
nvcc = /usr/local/cuda-10.2/bin/nvcc
cudalib = /usr/local/cuda-10.2/lib64/
tensorflow = /home/dpshah2/.conda/envs/PSRNet/lib/python2.7/site-packages/tensorflow/include
tf_lib = /home/dpshah2/.conda/envs/PSRNet/lib/python2.7/site-packages/tensorflow

all: depthestimate/tf_nndistance_so.so depthestimate/render_balls_so.so
.PHONY : all

depthestimate/tf_nndistance_so.so: depthestimate/tf_nndistance_g.cu.o depthestimate/tf_nndistance.cpp
g++ -std=c++11 depthestimate/tf_nndistance.cpp depthestimate/tf_nndistance_g.cu.o -o depthestimate/tf_nndistance_so.so -shared -fPIC -I $(tensorflow) -ltensorflow_framework -L $(tf_lib) -O2 -D_GLIBCXX_USE_CXX11_ABI=0

depthestimate/tf_nndistance_g.cu.o: depthestimate/tf_nndistance_g.cu
$(nvcc) -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++11 -c -o depthestimate/tf_nndistance_g.cu.o depthestimate/tf_nndistance_g.cu -I $(tensorflow) -DGOOGLE_CUDA=1 -x cu -Xcompiler -fPIC -O2

depthestimate/render_balls_so.so: depthestimate/render_balls_so.cpp
g++ -std=c++11 depthestimate/render_balls_so.cpp -o depthestimate/render_balls_so.so -shared -fPIC -O2 -D_GLIBCXX_USE_CXX11_ABI=0

(fastai-3.7) D:\git_hub\3D\PointSetGeneration\demo>python r2n2_128.py leg.png r2n2_128_v1.pkl
C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\h5py_init_.py:72: UserWarning: h5py is running against HDF5 1.10.2 when it was built against 1.10.3, this may cause problems
'{0}.{1}.{2}'.format(*version.hdf5_built_version_tuple)
curses is not supported on this machine (please install/reinstall curses for an optimal experience)
WARNING:tensorflow:From C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\tensorflow\python\framework\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Colocations handled automatically by placer.
WARNING:tensorflow:From C:\Users\Garima\Anaconda3\envs\fastai-3.7\lib\site-packages\tflearn\initializations.py:119: UniformUnitScaling.init (from tensorflow.python.ops.init_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.initializers.variance_scaling instead with distribution=uniform to get equivalent behavior.
2019-09-11 18:01:23.634506: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2
2019-09-11 18:01:23.828876: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1433] Found device 0 with properties:
name: GeForce GTX 1070 with Max-Q Design major: 6 minor: 1 memoryClockRate(GHz): 1.379
pciBusID: 0000:01:00.0
totalMemory: 8.00GiB freeMemory: 6.63GiB
2019-09-11 18:01:23.839396: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0
2019-09-11 18:01:24.375509: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-09-11 18:01:24.379886: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0
2019-09-11 18:01:24.382626: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N
2019-09-11 18:01:24.386412: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6380 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1070 with Max-Q Design, pci bus id: 0000:01:00.0, compute capability: 6.1)
Traceback (most recent call last):
File "r2n2_128.py", line 97, in
ret=run_image(model,img_in)
File "r2n2_128.py", line 86, in run_image
assert(img_in.shape==(HEIGHT,WIDTH,4))
AssertionError

When I executed Python view. py 1. png. txt, I got unable to connect to X server. How to solve it? Thank you for helping.

Has the function _nn_distance_grad in tf_nndistance.py been called? Is this computational gradient useful when updating?

In my code, I didn't use it, and my result is very bed：After training, the network weights are all close to zero, and this leads to the same for point clouds

If _nn_distance_grad has nothing to do with the result, what caused it?

Hope someone can answer for me, thank you

Hello, I read the paper and the experiment only compare the 1-view's result to 3D-R2N2.Could the model be used with Multiview?

Seems the rendered image and the point cloud are not aligned with the given data and the given data io API, is this supposed to be not aligned?

In this paper, the author's method compared with 3D-R2N2 by IoU. I want to know how to use it ? Can u help me?

Hi, is there a way to access the 3x3 calibration matrix with intrinsic parameters for rendered views? Thank you in advance!

Hi there,

I was trying to make this project.
But got the error
make: /usr/local/cuda-8.0/bin/nvcc: Command not found
is that because I was using anaconda?
should I replace the
nvcc = /usr/local/cuda-9.0/bin/nvcc
with my anaconda cuda address?
any way to solve it?

Thanks in advance

Since I had trouble with the interactive_crop myself, I thought to share this with anyone else who might come across it.

you can replace the
show_cropped=((cropped-(cropped/2*cropped_mask[:,:,None]))*(~overlay_mask))|(overlay*(overlay_mask))
in line 145 simply with
show_cropped=cropped*(~overlay_mask).

Hi there,

I have a question. Could you please explain what's the meaning of this code?

for i in xrange(FETCH_BATCH_SIZE):
    ptcloud[i]=((ptcloud[i]-[0.7,0.5,0.5])/0.4).dot(rotmat[i])+[1,0,0]

The ptcloud was first divided by 255, and then was subtracted by [0.7 0.5 0.5], what's the reason to do this? Thanks.

I have encountered an issue during training. It shows 'ConfigProto' object has no attribute 'name' in the terminal when I start training. Does anybody face similar problem and how to solve it. Thanks a lot

Hi!

In the paper you mentioned a model that you train to post process your point set to get a voxelized representation which you use to compare to R2N2.

For my use case I need voxels. Is there any chance you could release this model/code as well?

Thanks!

If you are looking for EMD of dense point clouds (with over 10,000 points) and large batch size, please check our implementation: https://github.com/Colin97/MSN-Point-Cloud-Completion

Python: 3.6
Ubuntu: 18.04
Tensorflow: 1.13.1

There is an error when loading the '.gz' files:
Traceback (most recent call last): File "/home/tifo-kj/anaconda3/envs/learn/lib/python3.6/threading.py", line 916, in _bootstrap_inner self.run() File "/home/tifo-kj/Documents/PointSetGeneration/depthestimate/BatchFetcher.py", line 71, in run self.queue.put(self.work(self.bno%300000)) File "/home/tifo-kj/Documents/PointSetGeneration/depthestimate/BatchFetcher.py", line 39, in work binfile=zlib.decompress(open(path,'r').read()) File "/home/tifo-kj/anaconda3/envs/learn/lib/python3.6/codecs.py", line 321, in decode (result, consumed) = self._buffer_decode(data, self.errors, final) UnicodeDecodeError: 'utf-8' codec can't decode byte 0x9c in position 1: invalid start byte

I tried some suggestions online, but failed. Could you please help? Thank you very mush!

please add a license