EyeQ_process.py

import fundus_prep as prep
import glob
import os
import cv2 as cv
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True

def process(image_list, save_path, extension):
    
    for image_path in image_list:
        dst_image = os.path.splitext(image_path.split('/')[-1])[0]+extension#'.png'
        dst_path = os.path.join(save_path, dst_image)
        if os.path.exists(dst_path):
            print('continue...')
            continue
        try:
            img = prep.imread(image_path)
            r_img, borders, mask = prep.process_without_gb(img)
            # r_img = cv.resize(r_img, (800, 800))
            prep.imwrite(dst_path, r_img)
            # mask = cv.resize(mask, (800, 800))
            # prep.imwrite(os.path.join('./original_mask', dst_image), mask)
        except:
            print(image_path)
            continue


if __name__ == "__main__":

    for i in range(3):
        image_list = glob.glob(os.path.join(f'/mnt/sda2/datasets/eyeq/test/{i}', '*.jpeg'))
        os.makedirs(f'/mnt/sda2/datasets/eyeq/test_preprocess/{i}', exist_ok = True)
        save_path = prep.fold_dir(f'/mnt/sda2/datasets/eyeq/test_preprocess/{i}')
        print(image_list[:10], save_path)
        process(image_list, save_path, ".jpg")

    #image_list = glob.glob(os.path.join('./original_img', '*.jpeg'))
    #save_path = prep.fold_dir('./original_crop')

    #process(image_list, save_path)