File size: 1,573 Bytes
5f3481c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 | import numpy as np
from PIL import Image
from tqdm import tqdm
import torch
import glob
import pdb
import os
from skimage.io import imsave
import argparse
parser = argparse.ArgumentParser(description="resize image using nearest-neighbor upsampling")
parser.add_argument("--img_dir", default='/mnt/session_space/home/lingzzha/EgoHOS/data/train/label', type=str)
parser.add_argument("--mask_tf", default=True, type=bool)
parser.add_argument("--max_scale", default=512, type=int)
args = parser.parse_args()
save_dir = args.img_dir+'_'+str(args.max_scale)
os.makedirs(save_dir, exist_ok = True)
for file in tqdm(glob.glob(args.img_dir + '/*')):
fname = file.split('/')[-1]
img = Image.open(file); w, h = img.size
if w > h:
aspect_ratio = h/w
if args.mask_tf:
img_resize = Image.open(file).resize((args.max_scale, int(args.max_scale * aspect_ratio)), resample = Image.NEAREST)
else:
img_resize = Image.open(file).resize((args.max_scale, int(args.max_scale * aspect_ratio)), resample = Image.BICUBIC)
else:
aspect_ratio = w/h
if args.mask_tf:
img_resize = Image.open(file).resize((int(args.max_scale * aspect_ratio), args.max_scale), resample = Image.NEAREST)
else:
img_resize = Image.open(file).resize((int(args.max_scale * aspect_ratio), args.max_scale), resample = Image.BICUBIC)
# if not fname.startswith('20210619'):
# img_resize = Image.fromarray(np.array(np.rot90(img_resize, 3)))
img_resize.save(os.path.join(save_dir, fname)) |