卷积神经网络

1.简单黑白边界检测

import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import paddle
from paddle.nn import Conv2D
from paddle.nn.initializer import Assign
%matplotlib inline

 
w = np.array([1, 0, -1], dtype='float32')
w = w.reshape([1, 1, 1, 3])
conv = Conv2D(in_channels=1, out_channels=1, kernel_size=[1, 3],
       weight_attr=paddle.ParamAttr(
          initializer=Assign(value=w)))

 
img = np.ones([50,50], dtype='float32')
img[:, 30:] = 0.
 
x = img.reshape([1,1,50,50])
 
x = paddle.to_tensor(x)
 
y = conv(x)
 
out = y.numpy()
f = plt.subplot(121)
f.set_title('input image', fontsize=15)
plt.imshow(img, cmap='gray')
f = plt.subplot(122)
f.set_title('output featuremap', fontsize=15)
plt.imshow(out.squeeze(), cmap='gray')
plt.show()
 
print(conv.weight)
 
print(conv.bias)

 

2.图像中物体边界检测

import matplotlib.pyplot as plt
from PIL import Image
import numpy as np
import paddle
from paddle.nn import Conv2D
from paddle.nn.initializer import Assign

img = Image.open('./work/images/section1/000000098520.jpg')
 
w = np.array([[-1, -1, -1], [-1, 8, -1], [-1, -1, -1]], dtype='float32') / 8
w = w.reshape([1, 1, 3, 3])
w = np.repeat(w, 3, axis=1)
 
conv = Conv2D(in_channels=3, out_channels=1, kernel_size=[3, 3],
              weight_attr=paddle.ParamAttr(
                  initializer=Assign(value=w)))

 
x = np.array(img).astype('float32')
 
x = np.transpose(x, (2, 0, 1))
 
x = x.reshape(1, 3, img.height, img.width)
x = paddle.to_tensor(x)
y = conv(x)
out = y.numpy()
plt.figure(figsize=(20, 10))
f = plt.subplot(121)
f.set_title('input image', fontsize=15)
plt.imshow(img)
f = plt.subplot(122)
f.set_title('output feature map', fontsize=15)
plt.imshow(out.squeeze(), cmap='gray')
plt.show()

 

3.图像均值模糊

import paddle
import matplotlib.pyplot as plt
from PIL import Image
import numpy as np
from paddle.nn import Conv2D
from paddle.nn.initializer import Assign
 
img = Image.open('./work/images/section1/000000355610.jpg').convert('L')
img = np.array(img)

 
w = np.ones([1, 1, 5, 5], dtype = 'float32')/25
conv = Conv2D(in_channels=1, out_channels=1, kernel_size=[5, 5],
        weight_attr=paddle.ParamAttr(
         initializer=Assign(value=w)))
x = img.astype('float32')
x = x.reshape(1,1,img.shape[0], img.shape[1])
x = paddle.to_tensor(x)
y = conv(x)
out = y.numpy()

plt.figure(figsize=(20, 12))
f = plt.subplot(121)
f.set_title('input image')
plt.imshow(img, cmap='gray')

f = plt.subplot(122)
f.set_title('output feature map')
out = out.squeeze()
plt.imshow(out, cmap='gray')

plt.show()

 学会使用Conv2D算子完成一个图像边界检测的任务，理解了算子是通过什么样的方法进行黑白边界的判断，还学会了使用卷积网络检测图片明暗分界处，检测物体的外形轮廓，观察输出特征图跟原图之间的对应关系，还学会了使用卷积核，是用当前像素跟它邻域内的像素取平均，这样可以使图像上噪声比较大的点变得更平滑。