Keras 使用 ResNet 模型進(jìn)行實(shí)時(shí)預(yù)測

ResNet是一個(gè)預(yù)訓(xùn)練模型。它使用 ImageNet 進(jìn)行訓(xùn)練。在 ImageNet 上預(yù)訓(xùn)練的 ResNet 模型權(quán)重。它具有以下語法:

keras.applications.resnet.ResNet50 (
   include_top = True,
   weights = 'imagenet',
   input_tensor = None,
   input_shape = None,
   pooling = None, 
   classes = 1000
)

• include_top 指的是網(wǎng)絡(luò)頂部的全連接層。
• weights 指的是 ImageNet 上的預(yù)訓(xùn)練。
• input_tensor 指用作模型的圖像輸入的可選的 Keras 張量。
• input_shape 指可選的形狀元組。此模型的默認(rèn)輸入大小為 224x224。
• clasees 指用于對(duì)圖像進(jìn)行分類的可選數(shù)量的類。

讓我們通過寫一個(gè)簡單的例子來理解模型：

第 1 步：導(dǎo)入模塊

加載如下指定的必要模塊:

>>> import PIL
>>> from keras.preprocessing.image import load_img 
>>> from keras.preprocessing.image import img_to_array 
>>> from keras.applications.imagenet_utils import decode_predictions 
>>> import matplotlib.pyplot as plt
>>> import numpy as np
>>> from keras.applications.resnet50 import ResNet50
>>> from keras.applications import resnet50

第 2 步：選擇一個(gè)輸入

選擇一個(gè)輸入圖像，Lotus，如下所示：

>>> filename = 'banana.jpg'
>>> ## load an image in PIL format
>>> original = load_img(filename, target_size = (224, 224))
>>> print('PIL image size',original.size)
PIL image size (224, 224)
>>> plt.imshow(original)
<matplotlib.image.AxesImage object at 0x1304756d8>
>>> plt.show()

在這里，我們加載了一個(gè)圖像(banana.jpg)并顯示了它。

第 3 步：將圖像轉(zhuǎn)換為 NumPy 數(shù)組

將輸入的 Banana 轉(zhuǎn)換為 NumPy 數(shù)組，以便將其傳遞到模型中以進(jìn)行預(yù)測。

>>> #convert the PIL image to a numpy array
>>> numpy_image = img_to_array(original)
>>> plt.imshow(np.uint8(numpy_image))
<matplotlib.image.AxesImage object at 0x130475ac8>
>>> print('numpy array size',numpy_image.shape)
numpy array size (224, 224, 3)
>>> # Convert the image / images into batch format
>>> image_batch = np.expand_dims(numpy_image, axis = 0)
>>> print('image batch size', image_batch.shape)
image batch size (1, 224, 224, 3)
>>>

第 4 步：模型預(yù)測

將輸入輸入模型以獲得預(yù)測

>>> prepare the image for the resnet50 model >>>
>>> processed_image = resnet50.preprocess_input(image_batch.copy())
>>> # create resnet model
>>>resnet_model = resnet50.ResNet50(weights = 'imagenet')
>>> Downloavding data from https://github.com/fchollet/deep-learning-models/releas
es/download/v0.2/resnet50_weights_tf_dim_ordering_tf_kernels.h5
102858752/102853048 [==============================] - 33s 0us/step
>>> # get the predicted probabilities for each class
>>> predictions = resnet_model.predict(processed_image)
>>> # convert the probabilities to class labels
>>> label = decode_predictions(predictions)
Downloading data from https://storage.googleapis.com/download.tensorflow.org/
data/imagenet_class_index.json
40960/35363 [==================================] - 0s 0us/step
>>> print(label)

輸出

[
   [
      ('n07753592', 'banana', 0.99229723), 
      ('n03532672', 'hook', 0.0014551596), 
      ('n03970156', 'plunger', 0.0010738898), 
      ('n07753113', 'fig', 0.0009359837) , 
      ('n03109150', 'corkscrew', 0.00028538404)
   ]
]

模型就可以正確地將圖像預(yù)測為 banana。