search

ankurhanda / sunrgbd-meta-data

train test labels for sunrgbd
170 stars 31 forks source link

What are the names for each class in 13-label version of SUNRGBD #1

Closed claudehang closed 5 years ago

claudehang commented 5 years ago

Where this meta data could be found? Many thanks!

ankurhanda commented 5 years ago

You can check my SceneNet as well as SceneNet RGB-D paper. This repo does provide train and test 13 class labels and match it against the papers. Once you have done it you can perhaps create a PR and I will merge it to the repo. Thanks.

irfanICMLL commented 5 years ago

What is the label 0 mean in the class37? The background class or the 'ignore' class?

ankurhanda commented 5 years ago

ignore

irfanICMLL commented 5 years ago

Thanks

loveis98 commented 5 years ago

I loaded sunrgbd_train_test_labels.tar.gz, but I didn't understand, how the labeled image pixels correspond to the classes? F.e., seg[0][0] gives me 0.06666667 - float value. Would you help me?

ankurhanda commented 5 years ago

Are you reading them into float variables directly? You should only see integers.

loveis98 commented 5 years ago

I read images by matplotlib.image.mpimg.imread

ankurhanda commented 5 years ago

I did this 

Python 3.6.4 (default, Mar  9 2018, 23:15:12) 
[GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.39.2)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from scipy import misc
>>> A = misc.imread('img-010333.png')
>>> A
array([[0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       ...,
       [5, 5, 5, ..., 5, 5, 5],
       [5, 5, 5, ..., 0, 0, 0],
       [5, 5, 5, ..., 0, 0, 0]], dtype=uint8)
>>> import numpy as np
>>> np.amax(A)
7
loveis98 commented 5 years ago

ok, thanks:)