摘要：import os from glob import glob from PIL import Image path = './' for i in range(４): path_ = path+ str(i+1) file_list = glob(os.path.join(path_,'*.png
阅读全文
摘要：# file_list = os.listdir(path) # # for i in range(len(file_list)): # print(file_list[i]) # name = file_list[i].split('.')[-2] # path_hr = path+'/'+fil
阅读全文
摘要：from util_ import calc_psnr_and_ssim import torch import numpy as np from glob import glob #import glob是错的 import os from PIL import Image from torchv
阅读全文
摘要：time_start = time.time() SR_left = net() torch.cuda.synchronize() time_end = time.time()
阅读全文
摘要：import os from PIL import Image path = './hr/' cnt =1 file_list = os.listdir(path) for i in range(len(file_list)): path_l = path+file_list[i]+'/hr0.pn
阅读全文
摘要：从float64(x)范围[0，1]到uint8(y)范围[0，255]的转换 y = (x*255).astype(np.uint8))
阅读全文
摘要：import os from PIL import Image path = './input' file_list = os.listdir(path) for i in range(len(file_list)): print(file_list[i]) import os from PIL i
阅读全文
摘要：input_list = sorted([os.path.join("/root/userfolder/cl/TTSR/dataset/CUFED/train/input/input", name) for name in os.listdir("/root/userfolder/cl/TTSR/d
阅读全文
摘要：import torch import cv2 as cv from PIL import Image import numpy as np a = cv.imread('woman.png') print(a.shape)#[H,W,C] print(a.dtype) c = Image.open
阅读全文
摘要：论文理解：https://blog.csdn.net/muyiyushan/article/details/89197083 https://blog.csdn.net/GracePro/article/details/100637822?utm_medium=distribute.pc_relev
阅读全文
摘要：参考链接：https://blog.csdn.net/lqz790192593/article/details/89762081https://www.cnblogs.com/qyh1499816254/p/11356887.html在DataLoader 中，iter(dataloader) 返回
阅读全文
摘要：PSNR：Peak Signal to Noise Ratio，峰值信噪比，单位是dB，数值越大表示失真越小。PSNR是最普遍和使用最为广泛的一种图像客观评价指标，然而它是基于对应像素点间的误差， 即基于误差敏感的图像质量评价。由于并未考虑到人眼的视觉特性（人眼对空间频率较低的对比差异敏感度较高，人
阅读全文
摘要：转载自：https://blog.csdn.net/NCU_wander/article/details/105529494 1、** list变量前面加星号，字典变量前面加两个星号；是将其作为参量进行传递，传递进去之后仍然保持其原字典或者list的原性质；最为常见的就是argparse模块中的参数
阅读全文
