1.Pytorch读取数据
由于本次赛题我们使用Pytorch框架讲解具体的解决方案,接下来将是解决赛题的第一步使用Pytorch读取赛题数据。 在Pytorch中数据是通过Dataset进行封装,并通过DataLoder进行并行读取。所以我们只需要重载一下数据读取的逻辑就可以完成数据的读取。
import os, sys, glob, shutil, json
import cv2
from PIL import Image
import numpy as np
import torch
from torch.utils.data.dataset import Dataset
import torchvision.transforms as transforms
class SVHNDataset(Dataset):
def __init__(self, img_path, img_label, transform=None):
self.img_path = img_path
self.img_label = img_label
if transform is not None:
self.transform = transform
else:
self.transform = None
def __getitem__(self, index):
img = Image.open(self.img_path[index]).convert('RGB')
if self.transform is not None:
img = self.transform(img)
# 原始SVHN中类别10为数字0
lbl = np.array(self.img_label[index], dtype=np.int)
lbl = list(lbl) + (5 - len(lbl)) * [10]
return img, torch.from_numpy(np.array(lbl[:5]))
def __len__(self):
return len(self.img_path)
train_path = glob.glob('../input/train/*.png')
train_path.sort()
train_json = json.load(open('../input/train.json'))
train_label = [train_json[x]['label'] for x in train_json]
data = SVHNDataset(train_path, train_label,
transforms.Compose([
# 缩放到固定尺寸
transforms.Resize((64, 128)),
# 随机颜色变换
transforms.ColorJitter(0.2, 0.2, 0.2),
# 加入随机旋转
transforms.RandomRotation(5),
# 将图片转换为pytorch 的tesntor
# transforms.ToTensor(),
# 对图像像素进行归一化
# transforms.Normalize([0.485,0.456,0.406],[0.229,0.224,0.225])
]))
接下来我们将在定义好的Dataset基础上构建DataLoder,你可以会问有了Dataset为什么还要有DataLoder?其实这两个是两个不同的概念,是为了实现不同的功能。
-
Dataset:对数据集的封装,提供索引方式的对数据样本进行读取
-
DataLoder:对Dataset进行封装,提供批量读取的迭代读取
import os, sys, glob, shutil, json import cv2 from PIL import Image import numpy as np import torch from torch.utils.data.dataset import Dataset import torchvision.transforms as transforms class SVHNDataset(Dataset): def __init__(self, img_path, img_label, transform=None): self.img_path = img_path self.img_label = img_label if transform is not None: self.transform = transform else: self.transform = None def __getitem__(self, index): img = Image.open(self.img_path[index]).convert('RGB') if self.transform is not None: img = self.transform(img) # 原始SVHN中类别10为数字0 lbl = np.array(self.img_label[index], dtype=np.int) lbl = list(lbl) + (5 - len(lbl)) * [10] return img, torch.from_numpy(np.array(lbl[:5])) def __len__(self): return len(self.img_path) train_path = glob.glob('../input/train/*.png') train_path.sort() train_json = json.load(open('../input/train.json')) train_label = [train_json[x]['label'] for x in train_json] train_loader = torch.utils.data.DataLoader( SVHNDataset(train_path, train_label, transforms.Compose([ transforms.Resize((64, 128)), transforms.ColorJitter(0.3, 0.3, 0.2), transforms.RandomRotation(5), transforms.ToTensor(), transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]) ])), batch_size=10, # 每批样本个数 shuffle=False, # 是否打乱顺序 num_workers=10, # 读取的线程个数 )
2.数据扩增
-
裁剪——Crop
中心裁剪:transforms.CenterCrop
随机裁剪:transforms.RandomCrop
随机长宽比裁剪:transforms.RandomResizedCrop
上下左右中心裁剪:transforms.FiveCrop
上下左右中心裁剪后翻转,transforms.TenCrop -
翻转和旋转——Flip and Rotation
依概率p水平翻转:transforms.RandomHorizontalFlip(p=0.5)
依概率p垂直翻转:transforms.RandomVerticalFlip(p=0.5)
随机旋转:transforms.RandomRotation -
图像变换
resize:transforms.Resize
标准化:transforms.Normalize
转为tensor,并归一化至[0-1]:transforms.ToTensor
填充:transforms.Pad
修改亮度、对比度和饱和度:transforms.ColorJitter
转灰度图:transforms.Grayscale
线性变换:transforms.LinearTransformation()
仿射变换:transforms.RandomAffine
依概率p转为灰度图:transforms.RandomGrayscale
将数据转换为PILImage:transforms.ToPILImage
transforms.Lambda:Apply a user-defined lambda as a transform. -
对transforms操作,使数据增强更灵活
transforms.RandomChoice(transforms), 从给定的一系列transforms中选一个进行操作
transforms.RandomApply(transforms, p=0.5),给一个transform加上概率,依概率进行操作
transforms.RandomOrder,将transforms中的操作随机打乱