File size: 7,821 Bytes
b213d84
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
# Copyright (c) Facebook, Inc. and its affiliates.
import json
import logging
import os

from detectron2.data import DatasetCatalog, MetadataCatalog
from detectron2.data.datasets.builtin_meta import CITYSCAPES_CATEGORIES
from detectron2.utils.file_io import PathManager

"""
This file contains functions to register the Cityscapes panoptic dataset to the DatasetCatalog.
"""


logger = logging.getLogger(__name__)


def get_cityscapes_panoptic_files(image_dir, gt_dir, json_info):
    files = []
    # scan through the directory
    cities = PathManager.ls(image_dir)
    logger.info(f"{len(cities)} cities found in '{image_dir}'.")
    image_dict = {}
    for city in cities:
        city_img_dir = os.path.join(image_dir, city)
        for basename in PathManager.ls(city_img_dir):
            image_file = os.path.join(city_img_dir, basename)

            suffix = "_leftImg8bit.png"
            assert basename.endswith(suffix), basename
            basename = os.path.basename(basename)[: -len(suffix)]

            image_dict[basename] = image_file

    for ann in json_info["annotations"]:
        image_file = image_dict.get(ann["image_id"], None)
        assert image_file is not None, "No image {} found for annotation {}".format(
            ann["image_id"], ann["file_name"]
        )
        label_file = os.path.join(gt_dir, ann["file_name"])
        segments_info = ann["segments_info"]

        files.append((image_file, label_file, segments_info))

    assert len(files), "No images found in {}".format(image_dir)
    assert PathManager.isfile(files[0][0]), files[0][0]
    assert PathManager.isfile(files[0][1]), files[0][1]
    return files


def load_cityscapes_panoptic(image_dir, gt_dir, gt_json, meta):
    """
    Args:
        image_dir (str): path to the raw dataset. e.g., "~/cityscapes/leftImg8bit/train".
        gt_dir (str): path to the raw annotations. e.g.,
            "~/cityscapes/gtFine/cityscapes_panoptic_train".
        gt_json (str): path to the json file. e.g.,
            "~/cityscapes/gtFine/cityscapes_panoptic_train.json".
        meta (dict): dictionary containing "thing_dataset_id_to_contiguous_id"
            and "stuff_dataset_id_to_contiguous_id" to map category ids to
            contiguous ids for training.

    Returns:
        list[dict]: a list of dicts in Detectron2 standard format. (See
        `Using Custom Datasets </tutorials/datasets.html>`_ )
    """

    def _convert_category_id(segment_info, meta):
        if segment_info["category_id"] in meta["thing_dataset_id_to_contiguous_id"]:
            segment_info["category_id"] = meta["thing_dataset_id_to_contiguous_id"][
                segment_info["category_id"]
            ]
        else:
            segment_info["category_id"] = meta["stuff_dataset_id_to_contiguous_id"][
                segment_info["category_id"]
            ]
        return segment_info

    assert os.path.exists(
        gt_json
    ), "Please run `python cityscapesscripts/preparation/createPanopticImgs.py` to generate label files."  # noqa
    with open(gt_json) as f:
        json_info = json.load(f)
    files = get_cityscapes_panoptic_files(image_dir, gt_dir, json_info)
    ret = []
    for image_file, label_file, segments_info in files:
        sem_label_file = (
            image_file.replace("leftImg8bit", "gtFine").split(".")[0] + "_labelTrainIds.png"
        )
        segments_info = [_convert_category_id(x, meta) for x in segments_info]
        ret.append(
            {
                "file_name": image_file,
                "image_id": "_".join(
                    os.path.splitext(os.path.basename(image_file))[0].split("_")[:3]
                ),
                "sem_seg_file_name": sem_label_file,
                "pan_seg_file_name": label_file,
                "segments_info": segments_info,
            }
        )
    assert len(ret), f"No images found in {image_dir}!"
    assert PathManager.isfile(
        ret[0]["sem_seg_file_name"]
    ), "Please generate labelTrainIds.png with cityscapesscripts/preparation/createTrainIdLabelImgs.py"  # noqa
    assert PathManager.isfile(
        ret[0]["pan_seg_file_name"]
    ), "Please generate panoptic annotation with python cityscapesscripts/preparation/createPanopticImgs.py"  # noqa
    return ret


_RAW_CITYSCAPES_PANOPTIC_SPLITS = {
    "cityscapes_fine_panoptic_train": (
        "cityscapes/leftImg8bit/train",
        "cityscapes/gtFine/cityscapes_panoptic_train",
        "cityscapes/gtFine/cityscapes_panoptic_train.json",
    ),
    "cityscapes_fine_panoptic_val": (
        "cityscapes/leftImg8bit/val",
        "cityscapes/gtFine/cityscapes_panoptic_val",
        "cityscapes/gtFine/cityscapes_panoptic_val.json",
    ),
    # "cityscapes_fine_panoptic_test": not supported yet
}


def register_all_cityscapes_panoptic(root):
    meta = {}
    # The following metadata maps contiguous id from [0, #thing categories +
    # #stuff categories) to their names and colors. We have to replica of the
    # same name and color under "thing_*" and "stuff_*" because the current
    # visualization function in D2 handles thing and class classes differently
    # due to some heuristic used in Panoptic FPN. We keep the same naming to
    # enable reusing existing visualization functions.
    thing_classes = [k["name"] for k in CITYSCAPES_CATEGORIES]
    thing_colors = [k["color"] for k in CITYSCAPES_CATEGORIES]
    stuff_classes = [k["name"] for k in CITYSCAPES_CATEGORIES]
    stuff_colors = [k["color"] for k in CITYSCAPES_CATEGORIES]

    meta["thing_classes"] = thing_classes
    meta["thing_colors"] = thing_colors
    meta["stuff_classes"] = stuff_classes
    meta["stuff_colors"] = stuff_colors

    # There are three types of ids in cityscapes panoptic segmentation:
    # (1) category id: like semantic segmentation, it is the class id for each
    #   pixel. Since there are some classes not used in evaluation, the category
    #   id is not always contiguous and thus we have two set of category ids:
    #       - original category id: category id in the original dataset, mainly
    #           used for evaluation.
    #       - contiguous category id: [0, #classes), in order to train the classifier
    # (2) instance id: this id is used to differentiate different instances from
    #   the same category. For "stuff" classes, the instance id is always 0; for
    #   "thing" classes, the instance id starts from 1 and 0 is reserved for
    #   ignored instances (e.g. crowd annotation).
    # (3) panoptic id: this is the compact id that encode both category and
    #   instance id by: category_id * 1000 + instance_id.
    thing_dataset_id_to_contiguous_id = {}
    stuff_dataset_id_to_contiguous_id = {}

    for k in CITYSCAPES_CATEGORIES:
        if k["isthing"] == 1:
            thing_dataset_id_to_contiguous_id[k["id"]] = k["trainId"]
        else:
            stuff_dataset_id_to_contiguous_id[k["id"]] = k["trainId"]

    meta["thing_dataset_id_to_contiguous_id"] = thing_dataset_id_to_contiguous_id
    meta["stuff_dataset_id_to_contiguous_id"] = stuff_dataset_id_to_contiguous_id

    for key, (image_dir, gt_dir, gt_json) in _RAW_CITYSCAPES_PANOPTIC_SPLITS.items():
        image_dir = os.path.join(root, image_dir)
        gt_dir = os.path.join(root, gt_dir)
        gt_json = os.path.join(root, gt_json)

        DatasetCatalog.register(
            key, lambda x=image_dir, y=gt_dir, z=gt_json: load_cityscapes_panoptic(x, y, z, meta)
        )
        MetadataCatalog.get(key).set(
            panoptic_root=gt_dir,
            image_root=image_dir,
            panoptic_json=gt_json,
            gt_dir=gt_dir.replace("cityscapes_panoptic_", ""),
            evaluator_type="cityscapes_panoptic_seg",
            ignore_label=255,
            label_divisor=1000,
            **meta,
        )