使用opencv及PIL在pytorch下读取目标检测coco数据集

使用opencv进行读取：

from torch.utils.data import Dataset,DataLoader,Subset,random_split
import re
from functools import reduce
from torch.utils.tensorboard import SummaryWriter as Writer
from torchvision import transforms,datasets
import torchvision as tv
from torch import nn
import torch.nn.functional as F
import time

t1=transforms.ToTensor()
t2=transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
t=transforms.Compose([t1,t2])

#数据集下载后为PIL格式：conda install git
TrainDataSet=datasets.CocoDetection(root=r"D:\coco2017\val2017\val2017",                               annFile=r"D:\coco2017\annotations_trainval2017\annotations\instances_val2017.json")
i=0
image,info=TrainDataSet[0]
image=np.array(image)
#调整为opencv使用的BGR
image = image[:, :, ::-1]
#为适配OPENCV转化为深拷贝
image=image.copy()
for annotation in info:
    x,y,width,height=annotation['bbox']
    #矩形函数必须转化为整数：
    x=int(x)
    y = int(y)
    width=int(width)
    height=int(height)
    cv2.rectangle(image, (x,y), (x+width, y+height), (0, 255, 0), 4)
cv2.imshow("emample",image)
cv2.waitKey()

使用PIL进行读取：

import os.path
from typing import Iterator
import numpy as np
import torch
import cv2
from PIL import Image,ImageDraw
from torch.utils.data import Dataset,DataLoader,Subset,random_split
import re
from functools import reduce
from torch.utils.tensorboard import SummaryWriter as Writer
from torchvision import transforms,datasets
import torchvision as tv
from torch import nn
import torch.nn.functional as F
import time

t1=transforms.ToTensor()
t2=transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
t=transforms.Compose([t1,t2])

#数据集下载后为PIL格式：conda install git
TrainDataSet=datasets.CocoDetection(root=r"D:\coco2017\val2017\val2017",
                                    annFile=r"D:\coco2017\annotations_trainval2017\annotations\instances_val2017.json")
i=0
image,info=TrainDataSet[0]
drawer=ImageDraw.ImageDraw(image)
for annotation in info:
    x,y,width,height=annotation['bbox']
    #注意打双括号，1个参数
    drawer.rectangle(((x,y), (x+width, y+height)))
image.show()