convert_images.py

import subprocess
import os
from pathlib import Path
from PIL import Image
import numpy as np

input_folder = Path("./original_images/")
output_folder = Path("./data/train/")
output_folder.mkdir(parents=True, exist_ok=True)
folders = os.listdir(input_folder)
for folder in folders:
    folder_path = input_folder.joinpath(folder)
    images = os.listdir(folder_path)
    for image in images:
        output = output_folder.joinpath(f'{folder}')
        output.mkdir(parents=True, exist_ok=True)

        output = output.joinpath(f'{folder}_{image}')
        output = str(output.absolute())

        input = folder_path.joinpath(image)
        input = str(input.absolute())

        if os.path.isfile(input) == False or '.json' in image:
            continue

        image_input = Image.open(input)

        background_color = (23, 35, 35, 255)
        new_color = (0, 0, 0, 255)
        data = np.array(image_input)
        data[(data == background_color).all(axis=-1)] = new_color
        image_input = Image.fromarray(data, 'RGBA')

        new_image = Image.new('RGB', image_input.size)
        new_image.paste(image_input, mask=image_input.split()[3])

        bbox = new_image.getbbox()
        new_image = new_image.crop(bbox)
        new_image.save(output)