Tensorflow numpy image reshape [grayscale images]

Question

I am trying to execute the Tensorflow "object_detection_tutorial.py" in jupyter notebook, with my trained neural network data but it throws a ValueError. The file mentioned above is part of Sentdexs tensorflow tutorial for object detection on youtube.

You can find it here: (https://www.youtube.com/watch?v=srPndLNMMpk&list=PLQVvvaa0QuDcNK5GeCQnxYnSSaar2tpku&index=6)

My Images are of Size: 490x704. So that would result in an 344960-array.

But it sais: ValueError: cannot reshape array of size 344960 into shape (490,704,3)

What am I doing wrong?

Code:

Imports

import numpy as np
import os
import six.moves.urllib as urllib
import sys
import tarfile
import tensorflow as tf
import zipfile

from collections import defaultdict
from io import StringIO
from matplotlib import pyplot as plt
from PIL import Image

Env setup

# This is needed to display the images.
%matplotlib inline

# This is needed since the notebook is stored in the object_detection folder.
sys.path.append("..")

Object detection imports

from utils import label_map_util

from utils import visualization_utils as vis_util

Variables

# What model to download.
MODEL_NAME = 'shard_graph'

# Path to frozen detection graph. This is the actual model that is used for the object detection.
PATH_TO_CKPT = MODEL_NAME + '/frozen_inference_graph.pb'

# List of the strings that is used to add correct label for each box.
PATH_TO_LABELS = os.path.join('training', 'object-detection.pbtxt')

NUM_CLASSES = 90

Load a (frozen) Tensorflow model into memory.

detection_graph = tf.Graph()
with detection_graph.as_default():
  od_graph_def = tf.GraphDef()
  with tf.gfile.GFile(PATH_TO_CKPT, 'rb') as fid:
    serialized_graph = fid.read()
    od_graph_def.ParseFromString(serialized_graph)
    tf.import_graph_def(od_graph_def, name='')

Loading label map

label_map = label_map_util.load_labelmap(PATH_TO_LABELS)
categories = label_map_util.convert_label_map_to_categories(label_map, max_num_classes=NUM_CLASSES, use_display_name=True)
category_index = label_map_util.create_category_index(categories)

Helper code

def load_image_into_numpy_array(image):
  (im_width, im_height) = image.size
  return np.array(image.getdata()).reshape(
      (im_height, im_width, 3)).astype(np.uint8)

Detection

# For the sake of simplicity we will use only 2 images:
# image1.jpg
# image2.jpg
# If you want to test the code with your images, just add path to the images to the TEST_IMAGE_PATHS.
PATH_TO_TEST_IMAGES_DIR = 'test_images'
TEST_IMAGE_PATHS = [ os.path.join(PATH_TO_TEST_IMAGES_DIR, 'frame_{}.png'.format(i)) for i in range(0, 2) ]

# Size, in inches, of the output images.
IMAGE_SIZE = (12, 8)

-

with detection_graph.as_default():
  with tf.Session(graph=detection_graph) as sess:
    # Definite input and output Tensors for detection_graph
    image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')
    # Each box represents a part of the image where a particular object was detected.
    detection_boxes = detection_graph.get_tensor_by_name('detection_boxes:0')
    # Each score represent how level of confidence for each of the objects.
    # Score is shown on the result image, together with the class label.
    detection_scores = detection_graph.get_tensor_by_name('detection_scores:0')
    detection_classes = detection_graph.get_tensor_by_name('detection_classes:0')
    num_detections = detection_graph.get_tensor_by_name('num_detections:0')
    for image_path in TEST_IMAGE_PATHS:
      image = Image.open(image_path)
      # the array based representation of the image will be used later in order to prepare the
      # result image with boxes and labels on it.
      image_np = load_image_into_numpy_array(image)
      # Expand dimensions since the model expects images to have shape: [1, None, None, 3]
      image_np_expanded = np.expand_dims(image_np, axis=0)
      # Actual detection.
      (boxes, scores, classes, num) = sess.run(
          [detection_boxes, detection_scores, detection_classes, num_detections],
          feed_dict={image_tensor: image_np_expanded})
      # Visualization of the results of a detection.
      vis_util.visualize_boxes_and_labels_on_image_array(
          image_np,
          np.squeeze(boxes),
          np.squeeze(classes).astype(np.int32),
          np.squeeze(scores),
          category_index,
          use_normalized_coordinates=True,
          line_thickness=8)
      plt.figure(figsize=IMAGE_SIZE)
      plt.imshow(image_np)

The last part of the script is throwing the Error:

----------------------------------------------------------------------
ValueError                           Traceback (most recent call last)
<ipython-input-62-7493eea60222> in <module>()
     14       # the array based representation of the image will be used later in order to prepare the
     15       # result image with boxes and labels on it.
---> 16       image_np = load_image_into_numpy_array(image)
     17       # Expand dimensions since the model expects images to have shape: [1, None, None, 3]
     18       image_np_expanded = np.expand_dims(image_np, axis=0)

<ipython-input-60-af094dcdd84a> in load_image_into_numpy_array(image)
      2   (im_width, im_height) = image.size
      3   return np.array(image.getdata()).reshape(
----> 4       (im_height, im_width, 3)).astype(np.uint8)

ValueError: cannot reshape array of size 344960 into shape (490,704,3)

Edit:

So I changed the last line in this function:

def load_image_into_numpy_array(image):
  (im_width, im_height) = image.size
  return np.array(image.getdata()).reshape(
      (im_height, im_width, 3)).astype(np.uint8)

to:

(im_height, im_width)).astype(np.uint8)

And the ValueError was solved. But now another ValueError connected to the array format is raised:

----------------------------------------------------------------------
ValueError                           Traceback (most recent call last)
<ipython-input-107-7493eea60222> in <module>()
     20       (boxes, scores, classes, num) = sess.run(
     21           [detection_boxes, detection_scores, detection_classes, num_detections],
---> 22           feed_dict={image_tensor: image_np_expanded})
     23       # Visualization of the results of a detection.
     24       vis_util.visualize_boxes_and_labels_on_image_array(

~/.local/lib/python3.6/site-packages/tensorflow/python/client/session.py in run(self, fetches, feed_dict, options, run_metadata)
    898     try:
    899       result = self._run(None, fetches, feed_dict, options_ptr,
--> 900                          run_metadata_ptr)
    901       if run_metadata:
    902         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

~/.local/lib/python3.6/site-packages/tensorflow/python/client/session.py in _run(self, handle, fetches, feed_dict, options, run_metadata)
   1109                              'which has shape %r' %
   1110                              (np_val.shape, subfeed_t.name,
-> 1111                               str(subfeed_t.get_shape())))
   1112           if not self.graph.is_feedable(subfeed_t):
   1113             raise ValueError('Tensor %s may not be fed.' % subfeed_t)

ValueError: Cannot feed value of shape (1, 490, 704) for Tensor 'image_tensor:0', which has shape '(?, ?, ?, 3)'

Does that mean that this tensorflow-model is not designed for grayscale images? Is there a way to make it work?

SOLUTION

Thanks to Matan Hugi it works just fine now. All I had to do is change this function to:

def load_image_into_numpy_array(image):
    # The function supports only grayscale images
    last_axis = -1
    dim_to_repeat = 2
    repeats = 3
    grscale_img_3dims = np.expand_dims(image, last_axis)
    training_image = np.repeat(grscale_img_3dims, repeats, dim_to_repeat).astype('uint8')
    assert len(training_image.shape) == 3
    assert training_image.shape[-1] == 3
    return training_image

Answer 1

Tensorflow expected input which is formated in NHWC format, which means: (BATCH, HEIGHT, WIDTH, CHANNELS).

Step 1 - Add last dimension:

last_axis = -1
grscale_img_3dims = np.expand_dims(image, last_axis)

Step 2 - Repeat the last dimension 3 times:

dim_to_repeat = 2
repeats = 3
np.repeat(grscale_img_3dims, repeats, dim_to_repeat)

So your function should be:

def load_image_into_numpy_array(image):
    # The function supports only grayscale images
    assert len(image.shape) == 2, "Not a grayscale input image" 
    last_axis = -1
    dim_to_repeat = 2
    repeats = 3
    grscale_img_3dims = np.expand_dims(image, last_axis)
    training_image = np.repeat(grscale_img_3dims, repeats, dim_to_repeat).astype('uint8')
    assert len(training_image.shape) == 3
    assert training_image.shape[-1] == 3
    return training_image