Incubator-mxnet: Invalid dimensions error when showing gray image in rec file

[albertyou2]

I can successfully show color images in .rec file by using this code blow:

show color images in .rec file

import mxnet as mx
import numpy as np
import matplotlib.pyplot as plt

import os

batch_size = 10
data_shape = (3, 100, 100)

rec_file = "/Users/zuekiWork/Documents/mxnet_jd/data/train_all_100.rec"
rec_list = "/Users/zuekiWork/Documents/mxnet_jd/data/train_all.lst"

rec_iter = mx.io.ImageRecordIter(
path_imgrec = rec_file,
path_imglist= rec_list,
data_shape = data_shape,
batch_size = batch_size,
shuffle = False,
flat = True)

X_test = rec_iter.getdata().asnumpy()

for i in range(batch_size):
plt.subplot(1,batch_size,i+1)
X=X_test[i]
print "before:", X.shape
X=np.swapaxes(X, 0, 2)
X=np.swapaxes(X, 0, 1)
print "after:", X.shape
#plt.imshow(X,cmap='Greys_r')
plt.imshow(X.astype(np.uint8))
plt.axis('off')
plt.show()

Then I modified the data_shape to (1,100,100) . When I show gray files in .rec file I get an error :.

Full exception log here :

....
Traceback (most recent call last):
File "show_img_in_rec.py", line 28, in
plt.imshow(X_test[0], cmap='Greys_r')
File "/Users/zuekiWork/anaconda2/lib/python2.7/site-packages/matplotlib/pyplot.py", line 3022, in imshow
*_kwargs)
File "/Users/zuekiWork/anaconda2/lib/python2.7/site-packages/matplotlib/init.py", line 1812, in inner
return func(ax, *args, *_kwargs)
File "/Users/zuekiWork/anaconda2/lib/python2.7/site-packages/matplotlib/axes/_axes.py", line 4947, in imshow
im.set_data(X)
File "/Users/zuekiWork/anaconda2/lib/python2.7/site-packages/matplotlib/image.py", line 453, in set_data
raise TypeError("Invalid dimensions for image data")
TypeError: Invalid dimensions for image data

could someone help me ?

[phunterlau]

This issue is closed due to lack of activity in the last 90 days. Feel free to reopen if this is still an active issue. Thanks!

[plant99]

The input of imshow can only be a 2D input - grayscale
or a 3D where the last dimension can have 3 ( RGB ) or 4 ( RGBA )
A good idea is to reshape the (x, y, 1) array to (x, y)