feat: Import yolo semantic segmentation labels

README.md

@@ -263,13 +263,19 @@ Use cases: image object detection
 
 ### YOLO to Label Studio converter 
 
-Usage:
+Object detection usage:
 
 ```
 label-studio-converter import yolo -i /yolo/root/directory -o ls-tasks.json
 ```
 
-Help:
+Semantic segmentation usage:
+
+```
+label-studio-converter import yolo-seg -i /yolo/root/directory -o ls-tasks.json
+```
+
+Help (same for semantic segmentation):
 
 ```
 label-studio-converter import yolo -h

label_studio_converter/imports/yolo_seg.py

@@ -0,0 +1,214 @@
+import os
+import json  # better to use "imports ujson as json" for the best performance
+
+import uuid
+import logging
+
+from PIL import Image
+from typing import Optional, Tuple
+from urllib.request import (
+    pathname2url,
+)  # for converting "+","*", etc. in file paths to appropriate urls
+
+from label_studio_converter.utils import ExpandFullPath
+from label_studio_converter.imports.label_config import generate_label_config
+
+logger = logging.getLogger('root')
+
+
+def convert_yolo_seg_to_ls(
+    input_dir,
+    out_file,
+    to_name='image',
+    from_name='label',
+    out_type="annotations",
+    image_root_url='/data/local-files/?d=',
+    image_ext='.jpg,.jpeg,.png',
+    image_dims: Optional[Tuple[int, int]] = None,
+):
+    """Convert YOLO labeling to Label Studio JSON
+    :param input_dir: directory with YOLO where images, labels, notes.json are located
+    :param out_file: output file with Label Studio JSON tasks
+    :param to_name: object name from Label Studio labeling config
+    :param from_name: control tag name from Label Studio labeling config
+    :param out_type: annotation type - "annotations" or "predictions"
+    :param image_root_url: root URL path where images will be hosted, e.g.: http://example.com/images
+    :param image_ext: image extension/s - single string or comma separated list to search, eg. .jpeg or .jpg, .png and so on.
+    :param image_dims: image dimensions - optional tuple of integers specifying the image width and height of *all* images in the dataset. Defaults to opening the image to determine it's width and height, which is slower. This should only be used in the special case where you dataset has uniform image dimesions.
+    """
+
+    tasks = []
+    logger.info('Reading YOLO notes and categories from %s', input_dir)
+
+    # build categories=>labels dict
+    notes_file = os.path.join(input_dir, 'classes.txt')
+    with open(notes_file) as f:
+        lines = [line.strip() for line in f.readlines()]
+    categories = {i: line for i, line in enumerate(lines)}
+    logger.info(f'Found {len(categories)} categories')
+
+    # generate and save labeling config
+    label_config_file = out_file.replace('.json', '') + '.label_config.xml'
+    generate_label_config(
+        categories,
+        {from_name: 'PolygonLabels'},
+        to_name,
+        from_name,
+        label_config_file,
+    )
+
+    # define directories
+    labels_dir = os.path.join(input_dir, 'labels')
+    images_dir = os.path.join(input_dir, 'images')
+    logger.info('Converting labels from %s', labels_dir)
+
+    # build array out of provided comma separated image_extns (str -> array)
+    image_ext = [x.strip() for x in image_ext.split(",")]
+    logger.info(f'image extensions->, {image_ext}')
+
+    # loop through images
+    for f in os.listdir(images_dir):
+        image_file_found_flag = False
+        for ext in image_ext:
+            if f.endswith(ext):
+                image_file = f
+                image_file_base = f[0 : -len(ext)]
+                image_file_found_flag = True
+                break
+        if not image_file_found_flag:
+            continue
+
+        image_root_url += '' if image_root_url.endswith('/') else '/'
+        task = {
+            "data": {
+                # eg. '../../foo+you.py' -> '../../foo%2Byou.py'
+                "image": image_root_url
+                + str(pathname2url(image_file))
+            }
+        }
+
+        # define coresponding label file and check existence
+        label_file = os.path.join(labels_dir, image_file_base + '.txt')
+
+        if os.path.exists(label_file):
+            task[out_type] = [
+                {
+                    "result": [],
+                    "ground_truth": False,
+                }
+            ]
+
+            # read image sizes
+            if image_dims is None:
+                # default to opening file if we aren't given image dims. slow!
+                with Image.open(os.path.join(images_dir, image_file)) as im:
+                    image_width, image_height = im.size
+            else:
+                image_width, image_height = image_dims
+
+            with open(label_file) as file:
+                # convert all polygons to Label Studio Results
+                lines = file.readlines()
+                for line in lines:
+                    label_id, *label_coords = line.split()
+                    label_coords = [float(coord) * 100 for coord in label_coords]
+                    vertices_xy = [label_coords[i: i + 2] for i in range(0, len(label_coords), 2)]
+
+                    item = {
+                        "id": uuid.uuid4().hex[0:10],
+                        "type": "polygonlabels",
+                        "value": {
+                            "points": vertices_xy,
+                            "closed": True,
+                            "polygonlabels": [categories[int(label_id)]],
+                        },
+                        "to_name": to_name,
+                        "from_name": from_name,
+                        "image_rotation": 0,
+                        "original_width": image_width,
+                        "original_height": image_height
+                    }
+                    task[out_type][0]["result"].append(item)
+
+        tasks.append(task)
+
+    if len(tasks) > 0:
+        logger.info('Saving Label Studio JSON to %s', out_file)
+        with open(out_file, 'w') as out:
+            json.dump(tasks, out)
+
+        print(
+            '\n'
+            f'  1. Create a new project in Label Studio\n'
+            f'  2. Use Labeling Config from "{label_config_file}"\n'
+            f'  3. Setup serving for images [e.g. you can use Local Storage (or others):\n'
+            f'     https://labelstud.io/guide/storage.html#Local-storage]\n'
+            f'  4. Import "{out_file}" to the project\n'
+        )
+    else:
+        logger.error('No labels converted')
+
+
+def add_parser(subparsers):
+    yolo_seg = subparsers.add_parser('yolo-seg')
+
+    yolo_seg.add_argument(
+        '-i',
+        '--input',
+        dest='input',
+        required=True,
+        help='directory with YOLO where images, labels, notes.json are located',
+        action=ExpandFullPath,
+    )
+    yolo_seg.add_argument(
+        '-o',
+        '--output',
+        dest='output',
+        help='output file with Label Studio JSON tasks',
+        default='output.json',
+        action=ExpandFullPath,
+    )
+    yolo_seg.add_argument(
+        '--to-name',
+        dest='to_name',
+        help='object name from Label Studio labeling config',
+        default='image',
+    )
+    yolo_seg.add_argument(
+        '--from-name',
+        dest='from_name',
+        help='control tag name from Label Studio labeling config',
+        default='label',
+    )
+    yolo_seg.add_argument(
+        '--out-type',
+        dest='out_type',
+        help='annotation type - "annotations" or "predictions"',
+        default='annotations',
+    )
+    yolo_seg.add_argument(
+        '--image-root-url',
+        dest='image_root_url',
+        help='root URL path where images will be hosted, e.g.: http://example.com/images',
+        default='/data/local-files/?d=',
+    )
+    yolo_seg.add_argument(
+        '--image-ext',
+        dest='image_ext',
+        help='image extension to search: .jpeg or .jpg, .png',
+        default='.jpg',
+    )
+    yolo_seg.add_argument(
+        '--image-dims',
+        dest='image_dims',
+        type=int,
+        nargs=2,
+        help=(
+            "optional tuple of integers specifying the image width and height of *all* "
+            "images in the dataset. Defaults to opening the image to determine it's width "
+            "and height, which is slower. This should only be used in the special "
+            "case where you dataset has uniform image dimesions. e.g. `--image-dims 600 800` "
+            "if all your images are of dimensions width=600, height=800"
+        ),
+        default=None,
+    )

label_studio_converter/main.py

@@ -6,7 +6,7 @@
 from label_studio_converter.converter import Converter, Format, FormatNotSupportedError
 from label_studio_converter.exports.csv import ExportToCSV
 from label_studio_converter.utils import ExpandFullPath
-from label_studio_converter.imports import yolo as import_yolo, coco as import_coco
+from label_studio_converter.imports import yolo as import_yolo, yolo_seg as import_yolo_seg, coco as import_coco
 
 logging.basicConfig(level=logging.INFO)
 
@@ -97,6 +97,7 @@ def get_all_args():
     )
     import_format = parser_import.add_subparsers(dest='import_format')
     import_yolo.add_parser(import_format)
+    import_yolo_seg.add_parser(import_format)
     import_coco.add_parser(import_format)
 
     return parser.parse_args()
@@ -166,6 +167,17 @@ def imports(args):
             image_root_url=args.image_root_url,
             image_ext=args.image_ext,
         )
+
+    elif args.import_format == 'yolo-seg':
+        import_yolo_seg.convert_yolo_seg_to_ls(
+            input_dir=args.input,
+            out_file=args.output,
+            to_name=args.to_name,
+            from_name=args.from_name,
+            out_type=args.out_type,
+            image_root_url=args.image_root_url,
+            image_ext=args.image_ext,
+        )
 
     elif args.import_format == 'coco':
         import_coco.convert_coco_to_ls(

tests/data/test_import_yolo_seg_data/classes.txt

@@ -0,0 +1,4 @@
+closed_door
+opened_door
+bus
+number

tests/data/test_import_yolo_seg_data/labels/img_1.txt

@@ -0,0 +1,4 @@
+2 0.0633721205070745 0.08960138500340462 0.05533468083300653 0.1070405807423223 0.046678976568625616 0.18461493420164576 0.037714140009088244 0.2970676791388046 0.03153149410595902 0.3938852830686579 0.028749303449550866 0.499121809079368 0.027203641973768566 0.5911286003915888 0.02844017115439441 0.6819326885494016 0.03400455246721071 0.7709327219756021 0.035241081647836556 0.8388377773093125 0.03678674312361886 0.8586823793570464 0.043587653617061 0.8683040045923114 0.10665064182897904 0.874317520364352 0.11190589084663888 0.8791283329819843 0.2559615403895497 0.8899526613716574 0.35921172697180764 0.8989729350297183 0.38270578140369876 0.9037837476473507 0.423202112069195 0.8977702318753102 0.4541153415848412 0.8911553645260656 0.4902838201181472 0.8779256298275763 0.5041947734001879 0.8779256298275763 0.5100682870081605 0.9031823960701466 0.5233609756998884 0.9200202402318602 0.5397449873431809 0.9152094276142279 0.5514920145591264 0.8941621224120857 0.5607659834138202 0.8610877856658626 0.5774591273522691 0.8496621056989855 0.6195011194935479 0.8334256131144762 0.6553604657316973 0.8213985815703948 0.6862736952473435 0.8099729016035176 0.718114321648459 0.8003512763682528 0.7450088313270711 0.7901282995557838 0.7802499129749075 0.781709377474927 0.812090539376023 0.7696823459308458 0.84640422413839 0.761263423849989 0.8525868700415196 0.7889255964013757 0.8658795587332472 0.8021553310998649 0.8748443952927846 0.7955404637506203 0.879790512015288 0.77028369750805 0.8835000995571655 0.7462296344198875 0.9147224613679683 0.7366080091846228 0.9379073835047026 0.728189087103766 0.9715598288456058 0.6957161019347466 0.9727963580262313 0.6229525610930557 0.9715598288456054 0.5435741529021202 0.9703232996649797 0.47020926048322503 0.9693959027795104 0.37519571128498397 0.968468505894041 0.282587568395559 0.9666137121231021 0.17795239396205292 0.9595036693345037 0.16472265926356364 0.9131338250610345 0.1569050887599109 0.8593448057038102 0.14608076037023787 0.8040101248708037 0.13645913513497296 0.7517667669893617 0.12743886147691208 0.6967412184515118 0.11726355755479129 0.6370786854863147 0.10643922916511822 0.5789618139969 0.09501354919824113 0.5242453977542065 0.08719597869458838 0.48776778692574413 0.08238516607695592 0.4383066197007104 0.07637165030491536 0.40028334739646565 0.07035813453287476 0.3585504875503434 0.06254056402922202 0.3189815537703164 0.0571283998343855 0.27415737097262954 0.0541216419483652 0.21855653267561792 0.0541216419483652 0.1765145405343392 0.05772975141158956 0.13509081298337347 0.06494597033803824 0.08995749789053015 0.0781757050365275
+0 0.3913081473705955 0.2077395452853236 0.42526224388285544 0.21093553828971323 0.46140692726687405 0.21306620029263962 0.49207514347149584 0.21519686229556606 0.4945910723225618 0.3550514436146151 0.4940932116698893 0.48773357743321344 0.49508893297523426 0.5874872984793128 0.49508893297523426 0.6882095022540152 0.496084654280579 0.7618141896278362 0.497080375585924 0.848009152473495 0.44928575292937056 0.8605994279453328 0.39501894178807545 0.8848114961604054 0.3935253598300581 0.7569717759848218 0.3945210811354031 0.6194472285232089 0.3940232204827305 0.5138826111054919 0.39153391721936837 0.3821689600154965 0.391036056566696 0.2814467562407942
+0 0.7565461408431807 0.23786503345366333 0.7874135013088714 0.23786503345366333 0.8217658863432692 0.24173896436807496 0.8222637469959416 0.34149268541417444 0.823757328953959 0.42381371734542156 0.824753050259304 0.5332522656775501 0.824753050259304 0.6572180549387222 0.8267444928699936 0.7482554314273956 0.7963749930569751 0.7608457068992333 0.7615247473699049 0.7695620514566595 0.7615247473699049 0.6562495722101194 0.7620226080225774 0.5690861266358577 0.7590354441065428 0.47223785377556693 0.7575418621485255 0.37054716727226167
+3 0.26465981600281774 0.1129307614638883 0.26117479143411076 0.22140082706741387 0.31593946322807825 0.22140082706741387 0.32191379106014745 0.10808834782087376

tests/data/test_import_yolo_seg_data/labels/img_2.txt

@@ -0,0 +1,4 @@
+2 0.8570140867160779 0.29644290501897036 0.7891087120622501 0.287059067307079 0.7298721086408257 0.27839706326533303 0.6706355052194016 0.2791188969354785 0.6118805001997775 0.27984073060562403 0.54493832234955 0.2863372336369334 0.4582506100255145 0.3180979151233352 0.3763788817194809 0.3476930955993005 0.30076793263685 0.3736791077245383 0.2434577228226265 0.3953341178289031 0.21648821232181542 0.4025524545303581 0.21070903150021308 0.45957731447185224 0.2092642362948125 0.5418663528684386 0.20733784268761168 0.6494195697201174 0.20733784268761168 0.715828267373503 0.23671534519742374 0.7461452815196137 0.25164489565322984 0.7504762835404866 0.25068169884962943 0.7721312936448517 0.2564608796712318 0.7721312936448517 0.259350470082033 0.785846133377616 0.26705604451083614 0.7930644700790708 0.2771696109486403 0.77718412933587 0.2805407997612417 0.7569727865717961 0.3368878127718647 0.7742967946552881 0.3797500705320823 0.7872898007179068 0.42550191870310106 0.7973954720999439 0.45247142920391215 0.8046138088013987 0.4563242164183137 0.837096323957946 0.4664377828561178 0.8616386687428929 0.488591309338927 0.8616386687428929 0.4982232773749309 0.8392618249683825 0.49918647417853124 0.8132758128431448 0.5304903702955441 0.8255469852356182 0.586355784904367 0.837096323957946 0.6301812394681849 0.8385399912982371 0.6985682125238131 0.8392618249683825 0.7236113294174232 0.8378181576280916 0.7332432974534272 0.8486456626802739 0.7650287919722402 0.854420332041438 0.7761055552136447 0.8378181576280916 0.7982590816964538 0.8349308229475095 0.8016302705090552 0.8125539791729992 0.842566134662072 0.8190504822043088 0.8748332275826853 0.8132758128431448 0.8859099908240898 0.8031701414611078 0.8878363844312903 0.7569727865717961 0.8690540467610828 0.7461452815196137 0.8700172435646834 0.6790147501960827 0.8690540467610828 0.6097187178621151 0.8685724483592826 0.5252641784550922 0.8661644563502816 0.4335913023466144 0.8637564643412806 0.37440094139468383
+0 0.4857017189281258 0.3902812821378847 0.5280823782865431 0.37656644240512027 0.5300087718937438 0.4104926249019585 0.5256743862775421 0.5332043488266928 0.5261559846793423 0.6025003811606602 0.5271191814829427 0.7194374357242305 0.5309719686973443 0.8082229771521263 0.4857017189281258 0.7973954720999439 0.4881097109371267 0.6833457522169557 0.488591309338927 0.559912194622076 0.48666491573172604 0.4718484868643257
+0 0.2949887518152476 0.43864413803763286 0.32051346711065803 0.4328694686764689 0.31810547510165704 0.5389790181878567 0.3161790814944563 0.6515850707305539 0.31762387669985687 0.7490326162001957 0.2940255550116472 0.7439797805091772 0.2925807598062466 0.6530287380708448 0.2940255550116472 0.575792535365277 0.293543956609847 0.49999999999999994
+3 0.5849109896989664 0.3029394080502798 0.5844293912971663 0.35563326597090095 0.6070645161817755 0.35707693331119195 0.608509311387176 0.30149574070998886

tests/data/test_import_yolo_seg_data/labels/img_3.txt

@@ -0,0 +1,4 @@
+3 0.2644538903363631 0.20499629894854596 0.26521803018928647 0.25593895581010573 0.29731190401206903 0.25729742665974725 0.29616569423268396 0.2043170635237252
+0 0.1535921104376025 0.2518635432611809 0.179572865436998 0.24235424731368976 0.18110114514284475 0.3272586754162893 0.17919079551053627 0.4338986371131542 0.18186528499576815 0.5452932467837648 0.1574128097022195 0.5364631862610945 0.15168176080529402 0.4325401662635127 0.15550246006991103 0.33608873593895955
+0 0.07717812514526295 0.2769952539795504 0.09398920190957767 0.27020289973134237 0.09246092220373088 0.374805155153745 0.09475334176250104 0.4705773500534773 0.08214503418926505 0.46989811462865644 0.07947054470403314 0.37548439057856586
+2 0.4233334785195612 0.1866569424783845 0.39544237388785725 0.1689968214330438 0.35799952109461086 0.1608459963351942 0.3136794096250539 0.15473287751180706 0.27012343800842037 0.15677058378626946 0.23764749425917608 0.16763835058340215 0.19485566249546588 0.18597770705356367 0.14671485176129195 0.20975094692229157 0.1043050899240435 0.23012800966691543 0.08061675448341825 0.24303348273851058 0.07221121610126088 0.26001436835903047 0.07182914617479919 0.35239038613465873 0.07297535595418428 0.4325401662635127 0.07373949580710767 0.4739735271775812 0.07443104237400336 0.5188132537471261 0.09754627292493608 0.5278538772514909 0.10216931903512264 0.5664819958610495 0.12019919886485014 0.5845632428697792 0.1317568141403165 0.5673038707250827 0.14100290636068957 0.5541538729005522 0.2117355118465437 0.5853851177338124 0.21635855795673029 0.6182601122951389 0.23300152395340176 0.6281226106635368 0.24964448995007338 0.624013236343371 0.2547298406712786 0.5919601166460776 0.25889058217044647 0.573878869637348 0.3698436888149235 0.5516882483084526 0.37307982109205406 0.5722351199092817 0.38787356864465106 0.5845632428697792 0.4045165346413226 0.5796319936855803 0.41283801763965833 0.5541538729005522 0.414687236083733 0.5410038750760214 0.44843547268809475 0.5286757521155241 0.4488977772991134 0.43498201761574345 0.44288781735587096 0.37169765308518987 0.43364172513549787 0.28868829181784034