Invalid image dimension while trying to visualize filters and features

[Rohit]

Hi

while I am trying to visualize my features and  filters, i am getting  errors:

    

File "netviz.py", line 53, in <module>

    vis_square(filters.transpose(0, 2, 3, 1))

  File "netviz.py", line 47, in vis_square

    plt.imshow(data)

  File "/usr/lib/pymodules/python2.7/matplotlib/pyplot.py", line 2892, in imshow

    imlim=imlim, resample=resample, url=url, **kwargs)

  File "/usr/lib/pymodules/python2.7/matplotlib/axes.py", line 7300, in imshow

    im.set_data(X)

  File "/usr/lib/pymodules/python2.7/matplotlib/image.py", line 429, in set_data

    raise TypeError("Invalid dimensions for image data")

TypeError: Invalid dimensions for image data

This is my python code and my image size is 288*288

import numpy as np

import matplotlib.pyplot as plt 

import matplotlib.cm as cm

import cv2 

plt.rcParams['figure.figsize'] = (6.0, 4.0)

import sys 

import caffe

MODEL_FILE = '/usr/share/caffe-master/models/bvlc_reference_caffenet/emotiondeploy.prototxt'

PRETRAINED = '/usr/share/caffe-master/models/bvlc_reference_caffenet/mydepthmodels/caffenet_train_iter_800.caffemodel'

IMAGE_FILE1 = '/home/iimtech7/depthdataset/train/neutral/neutral2.jpeg'

IMAGE_FILE2 = '/home/iimtech7/depthdataset/train/smile/smile2.jpeg'

blob = caffe.proto.caffe_pb2.BlobProto()

data = open( '/usr/share/caffe-master/examples/imagenet/mean.binaryproto' , 'rb' ).read()

blob.ParseFromString(data)

arr = np.array( caffe.io.blobproto_to_array(blob) )

out = arr[0]

net = caffe.Classifier(MODEL_FILE,PRETRAINED,mean=out,channel_swap=(2,1,0),raw_scale=255,image_dims=(288,288))

net2 = caffe.Classifier(MODEL_FILE,PRETRAINED,mean=out,channel_swap=(2,1,0),raw_scale=255,image_dims=(288,288))

input_image1 = caffe.io.load_image(IMAGE_FILE1)

input_image2 = caffe.io.load_image(IMAGE_FILE2)

# Display the test images

plt.subplot(1,4,1).imshow(input_image1),plt.title('neutral')

plt.subplot(1,4,2).imshow(input_image2),plt.title('side')

plt.show()

#prediction1 = net.predict([input_image1]) 

#prediction2 = net2.predict([input_image2])

#width=0.1

#plt.bar(np.arange(2),prediction1[0],width,color='blue',label='neutral')

#plt.bar(np.arange(2)+width,prediction2[0],width,color='green',label='side')

#plt.legend()

#plt.ylabel('Class Probability')

#plt.show()

def vis_square(data, padsize=1, padval=0):

    data -= data.min()

    data /= data.max()

    n = int(np.ceil(np.sqrt(data.shape[0])))

    padding = ((0, n ** 2 - data.shape[0]), (0, padsize), (0, padsize)) + ((0, 0),) * (data.ndim - 3)

    data = np.pad(data, padding, mode='constant', constant_values=(padval, padval))

    data = data.reshape((n, n) + data.shape[1:]).transpose((0, 2, 1, 3) + tuple(range(4, data.ndim + 1)))

    data = data.reshape((n * data.shape[1], n * data.shape[3]) + data.shape[4:])

    plt.imshow(data)

    plt.show()

    

plt.rcParams['figure.figsize'] = (50.0, 40.0)

filters = net.params['conv4'][0].data

vis_square(filters.transpose(0, 2, 3, 1)) 

feat = net.blobs['conv1'].data[0,:96]

plt.subplot(1,2,1),plt.title('neutral')

vis_square(feat, padval=1)

net.blobs['conv1'].data.shape

[Jan C Peters]

Well, as the error says: the dimensions (or "shape" in numpy speak) of "data" is one that imshow cannot understand. So try to "print data.shape" just before the imshow and see what shape it actually has. imshow can only handle [H, W, 3] or [H, W].

Jan