【无标题】

实现功能

读入图像文件夹以及掩码文件夹，通过文件名检测是否一一对应，检查无误后输入需要裁剪的区域大小，默认设定为256*256，仅支持裁剪正方形区域。通过鼠标点击确定左上顶点，会有蓝色框提示，确认裁剪区域后键入‘c’保存裁剪后的图像及掩码。

设置

在此处设置图片及掩码文件夹，以及输出文件夹，及裁剪区域大小

# Example usage:
image_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\image"
mask_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\label"
crop_size = int(input("Enter the crop size (e.g., 256 for 256x256): "))
out_img_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\image1"
out_mask_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\label1"
crop_images(image_folder, mask_folder, crop_size,out_img_folder,out_mask_folder)

在此处设置裁剪后文件命名规则

# Save cropped images with new names
new_img_name = 'crop_' + img_file
new_mask_name = 'crop_' + mask_file

功能示例

输入裁剪尺寸

通过点击图片确定裁剪区域

完整代码

import os
import cv2
import numpy as np

def crop_images(image_folder, mask_folder, crop_size,out_img_folder,out_mask_folder):
    # Ensure both folders exist
    if not os.path.exists(image_folder) or not os.path.exists(mask_folder):
        print("One or both folders do not exist.")
        return

    # Get list of files in the image folder
    image_files = sorted(os.listdir(image_folder))

    # Iterate through each image file
    for img_file in image_files:
        # Derive the corresponding mask file name
        mask_file = img_file.rsplit('.', 1)[0] + '_mask.' + img_file.rsplit('.', 1)[1]
        image_path = os.path.join(image_folder, img_file)
        mask_path = os.path.join(mask_folder, mask_file)

        # Read image and mask
        image = cv2.imread(image_path)
        mask = cv2.imread(mask_path, cv2.IMREAD_GRAYSCALE)

        if image is None or mask is None:
            print(f"Error reading {img_file} or {mask_file}. Skipping...")
            continue

        # Create a window and set mouse callback for cropping
        clone = image.copy()
        window_name = 'Select Crop Area'
        cv2.namedWindow(window_name)
        crop_rect = None
        cropping = False

        def mouse_callback(event, x, y, flags, param):
            nonlocal crop_rect, cropping
            if event == cv2.EVENT_LBUTTONDOWN:
                crop_rect = (x, y, x + crop_size, y + crop_size)
                cropping = True

        cv2.setMouseCallback(window_name, mouse_callback)

        while True:
            cv2.imshow(window_name, clone)
            key = cv2.waitKey(1) & 0xFF

            if cropping:
                cv2.rectangle(clone, (crop_rect[0], crop_rect[1]), (crop_rect[2], crop_rect[3]), (0, 255, 0), 2)
                cropping = False

            # Press 'c' to confirm crop
            if key == ord('c') and crop_rect:
                break

        cv2.destroyWindow(window_name)

        # Ensure crop_rect is valid and crop within the bounds of the image
        if crop_rect:
            x_start, y_start, x_end, y_end = crop_rect
            x_start, y_start = max(0, x_start), max(0, y_start)
            x_end, y_end = min(image.shape[1], x_end), min(image.shape[0], y_end)

            if x_end - x_start == crop_size and y_end - y_start == crop_size:
                cropped_image = image[y_start:y_end, x_start:x_end]
                cropped_mask = mask[y_start:y_end, x_start:x_end]

                # Save cropped images with new names
                new_img_name = 'crop_' + img_file
                new_mask_name = 'crop_' + mask_file
                cv2.imwrite(os.path.join(out_img_folder, new_img_name), cropped_image)
                cv2.imwrite(os.path.join(out_mask_folder, new_mask_name), cropped_mask)

                print(f"Cropped and saved {new_img_name} and {new_mask_name}")
            else:
                print(f"Invalid crop area for {img_file}. Skipping...")

    print("Processing complete.")

# Example usage:
image_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\image2"
mask_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\label2"
crop_size = int(input("Enter the crop size (e.g., 256 for 256x256): "))
out_img_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\image1"
out_mask_folder = r"E:\CAROTID\unet\Datasets\LEA\predict\label1"
crop_images(image_folder, mask_folder, crop_size,out_img_folder,out_mask_folder)