util/dataloader.py

from glob import glob

from torch.utils.data import Dataset
from PIL import Image
import math
import torch.nn.functional as F
import os

def prepadding(latent, factor=64):
    h, w = latent.size(2), latent.size(3)
    target_h = ((h - 1) // factor + 1) * factor
    target_w = ((w - 1) // factor + 1) * factor
    pad_h = (target_h - h) // 2
    pad_w = (target_w - w) // 2
    # 额外处理奇数padding的情况
    pad_h_extra = (target_h - h) % 2
    pad_w_extra = (target_w - w) % 2
    padded_latent = F.pad(latent, (pad_w, pad_w + pad_w_extra, pad_h, pad_h + pad_h_extra), mode='constant', value=0)   # 指定左、右、上、下的填充宽度
    # print("After padding: ", padded_latent.shape)
    return padded_latent, h, w

def crop_to_original_shape(blocks, ori_h, ori_w):
    _, _, padded_height, padded_width = blocks.shape
    start_h = (padded_height - ori_h) // 2
    end_h = start_h + ori_h
    start_w = (padded_width - ori_w) // 2
    end_w = start_w + ori_w
    cropped_blocks = blocks[:, :, start_h:end_h, start_w:end_w]
    # print("After cropping to original shape: ", cropped_blocks.shape)
    return cropped_blocks

class MSCOCO(Dataset):
    def __init__(self, root, transform, img_list=None):
        assert root[-1] == '/', "root to COCO dataset should end with \'/\', not {}.".format(
            root)

        if img_list:
            self.image_paths = []
            with open(img_list, 'r') as r:
                lines = r.read().splitlines()
                for line in lines:
                    self.image_paths.append(root + line)
        else:
            self.image_paths = sorted(glob(root + "*.jpg"))
        self.transform = transform

    def __getitem__(self, index):
        """
        Args:
            index (int): Index
        Returns:
            object: image.
        """
        img_path = self.image_paths[index]

        img = Image.open(img_path).convert('RGB')

        if self.transform is not None:
            img = self.transform(img)

        return img

    def __len__(self):
        return len(self.image_paths)

class MSCOCO_inference(Dataset):
    def __init__(self, root, transform, img_list=None):
        assert root[-1] == '/', "root to COCO dataset should end with \'/\', not {}.".format(
            root)

        if img_list:
            self.image_paths = []
            with open(img_list, 'r') as r:
                lines = r.read().splitlines()
                for line in lines:
                    self.image_paths.append(root + line)
        else:
            self.image_paths = sorted(glob(root + "*.jpg"))
        self.transform = transform

    def __getitem__(self, index):
        """
        Args:
            index (int): Index
        Returns:
            object: (image, filename).
        """
        img_path = self.image_paths[index]
        img = Image.open(img_path).convert('RGB')

        if self.transform is not None:
            img = self.transform(img)
        # print("img path=", img_path)
        filename = os.path.basename(img_path)  # 确保返回文件名字符串
        return img, filename

    def __len__(self):
        return len(self.image_paths)


class Kodak(Dataset):
    def __init__(self, root, transform):

        assert root[-1] == '/', "root to Kodak dataset should end with \'/\', not {}.".format(
            root)

        self.image_paths = sorted(glob(root + "*.png"))
        self.transform = transform

    def __getitem__(self, index):
        """
        Args:
            index (int): Index
        Returns:
            object: image.
        """
        img_path = self.image_paths[index]

        img = Image.open(img_path).convert('RGB')

        if self.transform is not None:
            img = self.transform(img)

        return img

    def __len__(self):
        return len(self.image_paths)