lisa_data/cocostuff.py

import numpy as np
import json
import os
from pycocotools import mask as maskUtils
from PIL import Image
from tqdm import tqdm
import glob

def singleMask2rle(mask):
    rle = maskUtils.encode(np.array(mask[:, :, None], order='F', dtype="uint8"))[0]
    rle["counts"] = rle["counts"].decode("utf-8")
    return rle


def init_cocostuff(base_image_dir):
    cocostuff_classes = []
    with open("../cocostuff_classes.txt") as f:
        for line in f.readlines()[1:]:
            cocostuff_classes.append(line.strip().split(": ")[-1])
    cocostuff_classes = np.array(cocostuff_classes)
    cocostuff_images = []

    cocostuff_labels = glob.glob(
        os.path.join(base_image_dir, "cocostuff", "train2017", "*.png")
    )
    cocostuff_images = [
        x.replace(".png", ".jpg").replace("cocostuff", "coco") for x in cocostuff_labels
    ]

    print("cocostuff: ", len(cocostuff_images))
    return cocostuff_classes, cocostuff_images, cocostuff_labels

base_image_dir = '/mnt/workspace/workgroup/yuanyq/code/LISA/dataset'
classes, images, labels = init_cocostuff(base_image_dir)
cocostuff_class2index = {
    c: i for i, c in enumerate(classes)
}

final_data = []
for idx in tqdm(range(len(images))):
    dic = {}
    image_path = images[idx]
    label_path = labels[idx]
    label = Image.open(label_path)
    label = np.array(label)
    for c, i in cocostuff_class2index.items():
        if "-" in c:
            label[label == i] = 255
            
    unique_label = np.unique(label).tolist()
    if 255 in unique_label:
        unique_label.remove(255)
    if len(unique_label) == 0:
        continue

    cats = []
    for class_id in unique_label:
        cats.append(classes[class_id])
    masks = []
    for class_id in unique_label:
        msk = label==class_id
        rle = singleMask2rle(msk)
        masks.append(rle)
    
    dic['image'] = image_path.replace('/mnt/workspace/workgroup/yuanyq/code/LISA/dataset/', '')
    dic['cat'] = cats
    dic['masks'] = masks
    final_data.append(dic)
    
print(len(final_data))
with open('cocostuff.json', 'w') as f:
    f.write(json.dumps(final_data))