使用python推荐的方式做图像resize

使用python推荐的方式做图像resize

最近，在学习吴恩达老师的dl课程作业时，遇到了一些问题，代码如下

## START CODE HERE ##
my_image = "my_image.jpg" # change this to the name of your image file 
my_label_y = [1] # the true class of your image (1 -> cat, 0 -> non-cat)
## END CODE HERE ##

fname = "images/" + my_image
image = np.array(ndimage.imread(fname, flatten=False))
my_image = scipy.misc.imresize(image, size=(num_px,num_px)).reshape((num_px*num_px*3,1))
my_predicted_image = predict(my_image, my_label_y, parameters)

plt.imshow(image)
print ("y = " + str(np.squeeze(my_predicted_image)) + ", your L-layer model predicts a \"" + classes[int(np.squeeze(my_predicted_image)),].decode("utf-8") +  "\" picture.")

我按照源代码运行，出现了以下几种错误

• AttributeError: module ‘scipy.ndimage’ has no attribute ‘imread’

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[63], line 7
      4 ## END CODE HERE ##
      6 fname = "images/" + my_image
----> 7 image = np.array(ndimage.imread(fname, flatten=False))
      8 my_image = scipy.misc.imresize(image, size=(num_px,num_px)).reshape((num_px*num_px*3,1))
      9 my_predicted_image = predict(my_image, my_label_y, parameters)

AttributeError: module 'scipy.ndimage' has no attribute 'imread'

• AttributeError: module ‘scipy.misc’ has no attribute ‘imresize’

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[65], line 9
      7 fname = "images/" + my_image
      8 image = Image.open(fname)
----> 9 my_image = scipy.misc.imresize(image, size=(num_px,num_px)).reshape((num_px*num_px*3,1))
     10 my_predicted_image = predict(my_image, my_label_y, parameters)
     12 plt.imshow(image)

AttributeError: module 'scipy.misc' has no attribute 'imresize'

在简单的查询官方文档后，得出是由于版本更新，一些方法被抛弃的原因。

那么接下来就是要用规范的方法将代码改动，以提高代码的实效性，首先分析代码做了什么

1. 打开图片（用数组的方式）
2. 将图片resize
3. 将数组reshape
4. 预测

分析完毕开始操作

打开图像方式

在百度上查看后，有的推荐降级，有的让使用其他的库，明显是比较粗糙的做法，于是查询官方找到目前推荐的打开图片方式，代码如下

from PIL import Image

my_image = "my_image.jpg" 
image = Image.open(my_image)

缩放图片

第二部就是缩放图片了，代码如下

image = image.resize((num_px,num_px))

完整代码

from PIL import Image

## START CODE HERE ##
my_image = "my_image.jpg" # change this to the name of your image file 
my_label_y = [1] # the true class of your image (1 -> cat, 0 -> non-cat)
## END CODE HERE ##

fname = "images/" + my_image
num_px = 64
image = Image.open(fname)
image = image.resize((num_px,num_px))
my_image = np.array(image).reshape((num_px*num_px*3,1))
print(my_image.shape)

my_predicted_image = predict(my_image, my_label_y, parameters)

plt.imshow(image)
print ("y = " + str(np.squeeze(my_predicted_image)) + ", your L-layer model predicts a \"" + classes[int(np.squeeze(my_predicted_image)),].decode("utf-8") +  "\" picture.")

运行结果

结论

遇到问题，先分析代码运行过程，然后合理查询文档，得到最优解决方案。既保证了代码的优美，又能掌握最新的知识，还能是代码有时效性，鲁棒性。