semantic segmentation code (from labelme) && semantic json labeling to instance labeling

ins.py

#!/usr/bin/env python

from __future__ import print_function

import argparse
import glob
import json
import os
import os.path as osp
import sys
from pprint import pprint
from tqdm import tqdm

import numpy as np
import PIL.Image

import labelme

def get_class_map():
    return {
        '_background_': 0,
        'Apple_porcelain': 0,
        'Beer': 0,
        'Book': 0,
        'Carbonated_drinks': 0,
        'Cup': 0,
        'Dispenser': 0,
        'Desk_lamp': 0,
        'Drugs': 0,
        'Glasses_case': 0,
        'Kettle': 0,
        'Milk': 0,
        'Mug': 0,
        'Metal_box': 0,
        'Nipper': 0,
        'Plate': 0,
        'Potato_chips': 0,
        'Seasoning_sauce': 0,
        'Spoon': 0,
        'Stick_snacks': 0,
        'Toy': 0,
        'Wash_product': 0,
    }
    
def main():
    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter
    )
    parser.add_argument('--input', help='input annotated directory')
    parser.add_argument('--output', help='output dataset directory')
    args = parser.parse_args()

    if osp.exists(args.output):
        print('Output directory already exists:', args.output)
    else:
        os.makedirs(args.output)
        os.makedirs(osp.join(args.output, 'JPEGImages'))
        os.makedirs(osp.join(args.output, 'SegmentationClass'))
        os.makedirs(osp.join(args.output, 'SegmentationClassPNG'))
        os.makedirs(osp.join(args.output, 'SegmentationClassVisualization'))
    print('Creating dataset:', args.output)

    colormap = labelme.utils.label_colormap(255)

    for label_file in tqdm(glob.glob(osp.join(args.input, '*.json'))):
        print('Generating dataset from:', label_file)
        with open(label_file) as f:
            base = osp.splitext(osp.basename(label_file))[0]
            out_img_file = osp.join(
                args.output, 'JPEGImages', base + '.jpg')
            out_lbl_file = osp.join(
                args.output, 'SegmentationClass', base + '.npy')
            out_png_file = osp.join(
                args.output, 'SegmentationClassPNG', base + '.png')
            out_viz_file = osp.join(
                args.output,
                'SegmentationClassVisualization',
                base + '.jpg',
            )

            data = json.load(f)
            
            shsh = data['shapes']
            class_map = get_class_map()
            
            for idx, sh in enumerate(shsh):
                label = sh['label']
                data['shapes'][idx]['label'] = label + str(class_map[label])
                class_map[label] += 1

            llk = [ds['label'] for ds in data['shapes']]
            llk.append('_background_')
            
            class_id = dict()
            
            for idx, ll in enumerate(llk):
                class_id[ll] = idx
            
            img_file = osp.join(osp.dirname(label_file), data['imagePath'])
            img = np.asarray(PIL.Image.open(img_file))
            PIL.Image.fromarray(img).save(out_img_file)

            lbl = labelme.utils.shapes_to_label(
                img_shape=img.shape,
                shapes=data['shapes'],
                label_name_to_value=class_id,
            )
            labelme.utils.lblsave(out_png_file, lbl)

            np.save(out_lbl_file, lbl)

            viz = labelme.utils.draw_label(
                lbl, img, tuple(llk), colormap=colormap)
            PIL.Image.fromarray(viz).save(out_viz_file)


if __name__ == '__main__':
    main()

seman.py

#!/usr/bin/env python

from __future__ import print_function

import argparse
import glob
import json
import os
import os.path as osp
import sys
from tqdm import tqdm

import numpy as np
import PIL.Image

import labelme


def main():
    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter
    )
    parser.add_argument('--input', help='input annotated directory')
    parser.add_argument('--output', help='output dataset directory')
    parser.add_argument('--labels', help='labels file', required=True)
    args = parser.parse_args()

    if osp.exists(args.output):
        print('Output directory already exists:', args.output)
    else:
        os.makedirs(args.output)
        os.makedirs(osp.join(args.output, 'JPEGImages'))
        os.makedirs(osp.join(args.output, 'SegmentationClass'))
        os.makedirs(osp.join(args.output, 'SegmentationClassPNG'))
        os.makedirs(osp.join(args.output, 'SegmentationClassVisualization'))
    print('Creating dataset:', args.output)

    class_names = []
    class_name_to_id = {}
    for i, line in enumerate(open(args.labels).readlines()):
        class_name = line.strip()
        class_name_to_id[class_name] = i
        
        class_names.append(class_name)
        
    print(class_name_to_id)
    class_names = tuple(class_names)
    print('class_names:', class_names)
    out_class_names_file = osp.join(args.output, 'class_names.txt')
    with open(out_class_names_file, 'w') as f:
        f.writelines('\n'.join(class_names))
    print('Saved class_names:', out_class_names_file)

    colormap = labelme.utils.label_colormap(255)

    for label_file in tqdm(glob.glob(osp.join(args.input, '*.json'))):
        print('Generating dataset from:', label_file)
        with open(label_file) as f:
            base = osp.splitext(osp.basename(label_file))[0]
            out_img_file = osp.join(
                args.output, 'JPEGImages', base + '.jpg')
            out_lbl_file = osp.join(
                args.output, 'SegmentationClass', base + '.npy')
            out_png_file = osp.join(
                args.output, 'SegmentationClassPNG', base + '.png')
            out_viz_file = osp.join(
                args.output,
                'SegmentationClassVisualization',
                base + '.jpg',
            )

            data = json.load(f)

            img_file = osp.join(osp.dirname(label_file), data['imagePath'])
            img = np.asarray(PIL.Image.open(img_file))
            PIL.Image.fromarray(img).save(out_img_file)

            lbl = labelme.utils.shapes_to_label(
                img_shape=img.shape,
                shapes=data['shapes'],
                label_name_to_value=class_name_to_id,
            )
            labelme.utils.lblsave(out_png_file, lbl)

            np.save(out_lbl_file, lbl)

            viz = labelme.utils.draw_label(
                lbl, img, class_names, colormap=colormap)
            PIL.Image.fromarray(viz).save(out_viz_file)


if __name__ == '__main__':
    main()