卷积池化模块介绍

import numpy as np
import torch
from torch import nn
import torch.nn as nn
from torch.autograd import Variable
import torch.nn.functional as F
from PIL import Image
import matplotlib.pyplot as plt
# % matplotlib inline
im =  Image.open('C:/Users/leo/Desktop/text/cat1.png').convert('L')#为灰度图像，每个像素用8个bit表示，0表示黑，255表示白，其他数字表示不同的灰度。

im =  np.array(im,   dtype='float32')#将灰度图片转化成矩阵
plt.imshow(im.astype('uint8'),cmap='gray')#cmap颜色设置
plt.rcParams['font.sans-serif']=['SimHei']
plt.rcParams['axes.unicode_minus'] = False#标题默认输出英文，如果输出中文，要对字体进行调整
plt.title('转化成灰度图片')
plt.show()

#将图片矩阵转化为pytorch tensor ,并适配卷积输入的要求
im =  torch.from_numpy(im.reshape((1,    1,im.shape[0],im.shape[1])))#shape[0],矩阵行数，shape[1],矩阵列数

conv1  =  nn.Conv2d(1,1,3,bias=False) #定义卷积，b为偏置
sobel_kernel   =  np.array([[-1, -1,    -1],   [-1,   8, -1],   [-1,   -1,    -1]],  dtype='float32')#定义轮廓检测算子

sobel_kernel = sobel_kernel.reshape((1,    1, 3, 3))#适配卷积的输入输出
conv1.weight.data  =  torch.from_numpy(sobel_kernel)#给卷积的核赋值
edge1  =  conv1(Variable(im))
edge1  =  edge1.data.squeeze().numpy()
plt.imshow(edge1,  cmap='gray')
plt.title('卷积后的图片')
plt.show()
pool1  =  nn.MaxPool2d(2,    2)
print('before  max    pool,  image  shape: {} x  {}'.format(im.shape[2],    im.shape[3]))#shape[0]输入通道数。shape[1]输出通道数，shape[2]高度，shape[3】宽度
small_im1  =  pool1(Variable(im))
small_im1  =  small_im1.data.squeeze().numpy()
print('after   max    pool,  image  shape: {} x  {} '.format(small_im1.shape[0],small_im1.shape[1]))
plt.imshow(small_im1,cmap='gray')
plt.show()

 

原图

操作代码后得到的图片