Question

我是TensorFlow和机器学习的新手。我正在尝试提取图像特征。我成功提取了1张图片的特征，但无法提取图片[] 这是脚本：

#read image
def readfile(pathdir):
paths = glob.glob(pathdir + "*.png")
img_data_list=[]
imgs = [cv2.imread(file,cv2.IMREAD_COLOR) for file in paths]


for img in imgs:
       img=cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
       img_resize=cv2.resize(img,(244,244),3)

       img_data_list.append(img_resize)

img_data = np.array(img_data_list)
np.expand_dims(img_data, axis=0)
#img_data = array(img_data).reshape(1, 244,244,3)
names = [os.path.basename(path) for path in paths]
with open(pathdir + "value.txt") as file:
    nvl= file.readlines()
nvl = {line.split()[0]:float(re.sub(r'\n', "", line.split()[1])) for line 
in nvl}
value =[]
for i in names:
  value.append(nvl[i])
return img_data, np.array(value), names
# extract features
def feature(train_imgs):
with tf.Session(config=tf.ConfigProto(gpu_options= 
(tf.GPUOptions(per_process_gpu_memory_fraction=0.7)))) as sess:
#with tf.device('/cpu:0'):
  #with tf.Session() as sess:
      images = tf.placeholder("float", [np.shape(train_imgs)[0], 224, 
224, 3])
     [bunch of code]
      return features

我收到以下错误：

ValueError                                Traceback (most recent call 
last)
 <ipython-input-23-b21a5d063db6> in feature(train_imgs)
 47               vgg.build(images)
 48           print("Extracting feature...")
 ---> 49           features = sess.run(vgg.fc6, feed_dict=feed_dict)
 50           print('FC6 feature: ', features)
 51           print('Number of input: ', len(features))

 /usr/local/lib/python3.6/dist- 
 packages/tensorflow/python/client/session.py in run(self, fetches, 
feed_dict, options, run_metadata)
927     try:
928       result = self._run(None, fetches, feed_dict, options_ptr,
--> 929                          run_metadata_ptr)
930       if run_metadata:
931         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

/usr/local/lib/python3.6/dist- 
packages/tensorflow/python/client/session.py in _run(self, handle, 
fetches, feed_dict, options, run_metadata)
1126                              'which has shape %r' %
1127                              (np_val.shape, subfeed_t.name,
-> 1128                               str(subfeed_t.get_shape())))
1129           if not self.graph.is_feedable(subfeed_t):
1130             raise ValueError('Tensor %s may not be fed.' % 
subfeed_t)

ValueError: Cannot feed value of shape (40, 244, 244) for Tensor  
Placeholder_4:0', which has shape '(40, 224, 224, 3)'

出了什么问题以及如何解决此问题。预先感谢！！

Answer 1

应用cv2.cvtColor（img，cv2.COLOR_BGR2GRAY）时，您将图像更改为灰度，仅获得一个颜色通道。您图像的形状现在是（？，244,244），并且您正在尝试填充形状（？，244,244,3）的占位符。如果要使用灰度图像，则应将图像集重塑为（？，244,244,1）并使用相同形状的占位符。

ValueError：无法为张量为'（40，224，224，3）'的张量'Placeholder_4：0'输入形状（40，244，244）的值

1 个答案: