|
|
import torch |
|
|
import os |
|
|
from PIL import Image |
|
|
import torchvision.transforms as transforms |
|
|
import matplotlib.pyplot as plt |
|
|
import numpy as np |
|
|
|
|
|
from utils.utils import generate_mask |
|
|
from PIL import Image |
|
|
|
|
|
class TrainDataset(torch.utils.data.Dataset): |
|
|
def __init__(self, data_path, transform=None): |
|
|
self.data = os.listdir(os.path.join(data_path, 'color')) |
|
|
self.data_path = data_path |
|
|
self.transform = transform |
|
|
self.ToTensor = transforms.ToTensor() |
|
|
|
|
|
|
|
|
self.bw_directory = os.path.join(data_path, 'bw') |
|
|
if not os.path.exists(self.bw_directory): |
|
|
os.makedirs(self.bw_directory) |
|
|
|
|
|
def __len__(self): |
|
|
return len(self.data) |
|
|
|
|
|
def __getitem__(self, idx): |
|
|
image_name = self.data[idx] |
|
|
|
|
|
color_img = Image.open(os.path.join(self.data_path, 'color', image_name)).convert('RGB') |
|
|
|
|
|
bw_name = "bw_" + image_name |
|
|
dfm_name = 'dfm_' + image_name |
|
|
|
|
|
bw_img = np.expand_dims(np.array(Image.open(os.path.join(self.bw_directory, bw_name)).convert('L')), 2) |
|
|
dfm_img = np.expand_dims(np.array(Image.open(os.path.join(self.bw_directory, dfm_name)).convert('L')), 2) |
|
|
|
|
|
bw_img = np.concatenate([bw_img, dfm_img], axis=2) |
|
|
|
|
|
if self.transform: |
|
|
result = self.transform(image=np.array(color_img), mask=bw_img) |
|
|
color_img = result['image'] |
|
|
bw_img = result['mask'] |
|
|
|
|
|
dfm_img = bw_img[:, :, 1] |
|
|
bw_img = bw_img[:, :, 0] |
|
|
|
|
|
color_img = self.ToTensor(color_img) |
|
|
bw_img = self.ToTensor(bw_img) |
|
|
dfm_img = self.ToTensor(dfm_img) |
|
|
|
|
|
color_img = (color_img - 0.5) / 0.5 |
|
|
|
|
|
mask = generate_mask(bw_img.shape[1], bw_img.shape[2]) |
|
|
hint = torch.cat((color_img * mask, mask), 0) |
|
|
|
|
|
return bw_img, color_img, hint, dfm_img |
|
|
|
|
|
class FineTuningDataset(torch.utils.data.Dataset): |
|
|
def __init__(self, data_path, transform=None): |
|
|
self.data = [x for x in os.listdir(os.path.join(data_path, 'real_manga')) if x.find('dfm_') == -1] |
|
|
self.color_data = [x for x in os.listdir(os.path.join(data_path, 'color'))] * 100 |
|
|
self.data_path = data_path |
|
|
self.transform = transform |
|
|
|
|
|
|
|
|
self.bw_directory = os.path.join(data_path, 'bw') |
|
|
if not os.path.exists(self.bw_directory): |
|
|
os.makedirs(self.bw_directory) |
|
|
|
|
|
np.random.shuffle(self.color_data) |
|
|
self.ToTensor = transforms.ToTensor() |
|
|
|
|
|
def __len__(self): |
|
|
return len(self.data) |
|
|
|
|
|
def __getitem__(self, idx): |
|
|
color_img = Image.open(os.path.join(self.data_path, 'color', self.color_data[idx])).convert('RGB') |
|
|
bw_name = "bw_" + self.data[idx] |
|
|
dfm_name = "dfm_" + self.data[idx] |
|
|
|
|
|
bw_img = np.expand_dims(np.array(Image.open(os.path.join(self.bw_directory, bw_name)).convert('L')), 2) |
|
|
dfm_img = np.expand_dims(np.array(Image.open(os.path.join(self.bw_directory, dfm_name)).convert('L')), 2) |
|
|
|
|
|
if self.transform: |
|
|
result = self.transform(image=np.array(color_img)) |
|
|
color_img = result['image'] |
|
|
|
|
|
result = self.transform(image=bw_img, mask=dfm_img) |
|
|
bw_img = result['image'] |
|
|
dfm_img = result['mask'] |
|
|
|
|
|
color_img = self.ToTensor(color_img) |
|
|
bw_img = self.ToTensor(bw_img) |
|
|
dfm_img = self.ToTensor(dfm_img) |
|
|
|
|
|
color_img = (color_img - 0.5) / 0.5 |
|
|
|
|
|
return bw_img, dfm_img, color_img |
|
|
|
