Resize the images in data set and write their names to the text file.

Image processing.py

# Creates txt file containing file names from the data set located in DATASET_DIR and
# crops the images according to IMAGE_WIDTH and IMAGE_HEIGHT

import os
import numpy
import shutil
from PIL import Image

# Resize images to 128x128, channel 3 because colored
IMAGE_WIDTH, IMAGE_HEIGHT = 128, 128
IMAGE_CHANNELS = 3

DRIVE_DIR = '/content/drive/My Drive/Colab Notebooks/datasets/sperm_datasets/SpermDataset/' # Google Drive path
DATASET_DIR = DRIVE_DIR + 'My_dataset/' # dataset path

DATA_TXT_NAME = 'data.txt' # dataset txt name
CROPPED_FOLDER = 'cropped/' # where cropped images will be stored

TXT_FILE = open(DATA_TXT_NAME, "w")
TXT_PATH = DRIVE_DIR + DATA_TXT_NAME

cropped_full_path = os.path.join(DRIVE_DIR, CROPPED_FOLDER)
save_dir = cropped_full_path

if not os.path.exists(cropped_full_path):
  os.mkdir(cropped_full_path)
  print(f"{CROPPED_FOLDER} created.")
else:
  print(f"{CROPPED_FOLDER} exists.")

# write data names to txt file
for filename in os.listdir(DATASET_DIR):
  filename = filename + '.jpg'
  print(filename)
  TXT_FILE.write(filename)
  TXT_FILE.write("\n")
TXT_FILE.close()

shutil.move(DATA_TXT_NAME, TXT_PATH) # move file to drive

# Iterating over the images inside the directory and resizing them using
# Pillow's resize method
print('resizing and saving...')

for filename in os.listdir(DATASET_DIR):
  path = os.path.join(DATASET_DIR, filename)
  if not (os.path.isdir(path)):
    image = Image.open(path).resize((IMAGE_WIDTH, IMAGE_HEIGHT), Image.ANTIALIAS)
    new_file_name = os.path.join(save_dir, filename)
    new_file_name = new_file_name + '.jpg'
    print(new_file_name)
    image.save(new_file_name)