목록Wireless Comm./Python (24)
UOMOP
send img with contour
import cv2 import math import time import torch import random import torchvision import numpy as np from PIL import Image import torch.nn as nn from numpy import sqrt from tqdm import trange import torch.optim as optim import torch.nn.functional as F import matplotlib.pyplot as plt import torchvision.transforms as tr from torchvision import datasets from sklearn.preprocessing import StandardScal..
trainloader image 확인
transf = tr.Compose([tr.ToTensor(), tr.Grayscale(num_output_channels=1) ]) trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transf) #위에 전처리 trainloader = DataLoader(trainset, batch_size=50, shuffle=True) for data in trainloader : input = data[0] print(type(input)) print(input.shape) plt.imshow(input[0][0]) break
New : add noise to image using Norm2
def add_noise(signal, snr_db): ''' signal: np.ndarray snr: float returns -> np.ndarray ''' if type(signal[0][0]) != "numpy.float64" : signal = signal.astype(float) sig_flattend = signal.flatten().tolist() Norm2 = np.linalg.norm(sig_flattend, axis = 0, ord = 2) signal *= signal * Norm2 snr = 10.0 ** (snr_db / 10.0) # Generate the noise as you did noise = acoustics.generator.white(signal.size).res..
Is the high frequency component more robust to noise? (MSE Ver.)
#################################################################################### SNR_length = 30 filtering_size = 30 #################################################################################### import acoustics def add_AWGN_to_img(img, SNRdB) : height, width = img.shape noise = acoustics.generator.white(img.size).reshape(img.shape) SNR = 10.0**(SNRdB/10.0) current_SNR = np.mean(img) ..
Is the high frequency component more robust to noise? (PSNR Ver.)
#################################################################################### SNR_length = 30 filtering_size = 30 #################################################################################### import acoustics def add_AWGN_to_img(img, SNRdB) : height, width = img.shape noise = acoustics.generator.white(img.size).reshape(img.shape) SNR = 10.0**(SNRdB/10.0) current_SNR = np.mean(img) ..
High freq. vs Low freq. vs Original
#################################################################################### SNR_length = 30 filtering_size = 50 #################################################################################### import acoustics def add_AWGN_to_img(img, SNRdB) : height, width = img.shape noise = acoustics.generator.white(img.size).reshape(img.shape) SNR = 10.0**(SNRdB/10.0) current_SNR = np.mean(img) ..