卷积
卷积运算:卷积核在输入信号(图像)上滑动,相应位置上进行乘加
卷积核:又称为滤波器,过滤器,可认为是某种模式,某种特征
卷积过程类似于一个模板去图像上寻找与它相似的区域,与卷积核模式越相似,激活值越高,从而实现特征提取
AlexNet 卷积核可视化,发现卷积核学习到的是边缘,条纹,色彩这些细节模式
卷积维度:一般情况下,卷积核在几个维度上滑动,就是几维卷积
nn.Conv2d
功能:对多个二维信号进行二维卷积
主要参数:
in_channels 输入通道数
out_channels 输出通道数,等价于卷积核个数
kernel_size 卷积核尺寸
stride 步长
padding 填充个数
dilation 空洞卷积大小
groups 分组卷积设置
bias 偏置
尺寸计算:
简化版
完整版
转置卷积
转置卷积又称为反卷积(Deconvolution)和部分跨越卷积(Fractionally-strided Convolution),用于对图像进行上采样(UpSample)
nn.ConvTranspose2d
功能 转置卷积实现上采样
参数:
in_channels 输入通道数
out_channels 输出通道数
kernel_size 卷积核尺寸
stride 步长
padding 填充个数
dilation 空洞卷积大小
groups 分组卷积设置
bias 偏置
# -*- coding: utf-8 -*-
import os
import torch.nn as nn
from PIL import Image
from torchvision import transforms
from matplotlib import pyplot as plt
from tools.common_tools import transform_invert, set_seed
set_seed(5) # 设置随机种子
# ================================= load img ==================================
# 读取图片
path_img = os.path.join(os.path.dirname(os.path.abspath(__file__)), "lena.png")
img = Image.open(path_img).convert('RGB') # 0~255
# convert to tensor
# 转换成张量的形式
img_transform = transforms.Compose([transforms.ToTensor()])
img_tensor = img_transform(img)
img_tensor.unsqueeze_(dim=0) # C*H*W to B*C*H*W
# ================================= create convolution layer ==================================
# ================ 2d
# flag = 1
flag = 0
if flag:
conv_layer = nn.Conv2d(3, 1, 3) # input:(i, o, size) weights:(o, i , h, w)
nn.init.xavier_normal_(conv_layer.weight.data)
# calculation
img_conv = conv_layer(img_tensor)
# ================ transposed
flag = 1
# flag = 0
if flag:
# stride 决定了整个上采样的大小
conv_layer = nn.ConvTranspose2d(3, 1, 3, stride=2) # input:(i, o, size)
nn.init.xavier_normal_(conv_layer.weight.data)
# calculation
img_conv = conv_layer(img_tensor)
# ================================= visualization ==================================
print("卷积前尺寸:{}\n卷积后尺寸:{}".format(img_tensor.shape, img_conv.shape))
img_conv = transform_invert(img_conv[0, 0:1, ...], img_transform)
img_raw = transform_invert(img_tensor.squeeze(), img_transform)
plt.subplot(122).imshow(img_conv, cmap='gray')
plt.subplot(121).imshow(img_raw)
plt.show()