Upload 10 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+example.png filter=lfs diff=lfs merge=lfs -text

class_names.txt

@@ -0,0 +1 @@
+court

element.yaml

@@ -0,0 +1,6 @@
+version: 0.1.0
+element_type: PitchCalib
+main: main.py
+source: https://universe.roboflow.com/roboflow-jvuqo/basketball-court-detection-2
+objects:
+  - court

environment.json

@@ -0,0 +1 @@
+{"BATCH_SIZE": 8, "CACHE_PATH": "/train/cache", "DATASET_ID": "tT8apWsPavyvfZNc4fgN", "DATASET_LINK": "https://app.roboflow.com/ds/0066reL7dH?key=6NrcEsI7rH", "DATASET_OWNER": "GGqN5h3Ag5Msh1nOv8tjZvWDZpv1", "DATASET_VERSION_ID": "19", "ENDPOINT": "tT8apWsPavyvfZNc4fgN/19", "MODEL_NAME": "yolo11m-pose", "PREPROCESSING": "{\"auto-orient\":{\"enabled\":true},\"resize\":{\"enabled\":true,\"format\":\"Stretch to\",\"width\":\"640\",\"height\":\"640\"}}", "PROJECT": "roboflow-platform", "RESOLUTION": 640, "TRAINING_TIME": "2678400", "UID": "GGqN5h3Ag5Msh1nOv8tjZvWDZpv1", "COLORS": {"court": "#C7FC00"}}

keypoints_metadata.json

@@ -0,0 +1 @@
+[{"object_class": "court", "object_class_id": 0, "keypoints": {"0": "01", "1": "02", "2": "04", "3": "05", "4": "07", "5": "08", "6": "09", "7": "10", "8": "11", "9": "12", "10": "13", "11": "14", "12": "15", "13": "16", "14": "17", "15": "19", "16": "21", "17": "23", "18": "25", "19": "26", "20": "27", "21": "28", "22": "29", "23": "30", "24": "31", "25": "32", "26": "33", "27": "34", "28": "35", "29": "37", "30": "38", "31": "40", "32": "41"}}]

main.py

@@ -0,0 +1,105 @@
+import argparse
+import json
+import sys
+from io import BytesIO
+from pathlib import Path
+from typing import Any, Dict, List
+
+import numpy as np
+from PIL import Image
+from ultralytics import YOLO
+
+
+def load_keypoint_names(base_dir: Path) -> dict[int, str]:
+    metadata_path = base_dir / "keypoints_metadata.json"
+    if not metadata_path.exists():
+        return {}
+    try:
+        data = json.loads(metadata_path.read_text())
+    except Exception:
+        return {}
+    if not data:
+        return {}
+    keypoints = data[0].get("keypoints", {})
+    names: dict[int, str] = {}
+    for idx_str, label in keypoints.items():
+        try:
+            idx = int(idx_str)
+        except ValueError:
+            continue
+        names[idx] = str(label)
+    return names
+
+
+def load_image(frame: Any, base_dir: Path) -> Image.Image:
+    if isinstance(frame, (bytes, bytearray, memoryview)):
+        return Image.open(BytesIO(frame)).convert("RGB")
+
+    path = Path(str(frame))
+    if not path.is_absolute():
+        path = (Path.cwd() / path).resolve()
+        if not path.exists():
+            candidate = (base_dir / str(frame)).resolve()
+            if candidate.exists():
+                path = candidate
+    return Image.open(path).convert("RGB")
+
+
+def load_model(*_args: Any, **_kwargs: Any):
+    base_dir = Path(__file__).resolve().parent
+    model_path = base_dir / "weights.onnx"
+    if not model_path.exists():
+        return None
+    return YOLO(str(model_path), task="pose")
+
+
+def run_model(model, frame: "np.ndarray") -> List[Dict[str, Any]]:
+    base_dir = Path(__file__).resolve().parent
+    keypoint_names = load_keypoint_names(base_dir)
+    image = Image.fromarray(frame)
+    results = model(image)
+    outputs: List[Dict[str, Any]] = []
+    result = results[0]
+    if result.keypoints is None:
+        return outputs
+
+    keypoints_xy = result.keypoints.xy
+    for det_idx, points in enumerate(keypoints_xy):
+        point_items = []
+        for idx, xy in enumerate(points.tolist()):
+            label = keypoint_names.get(idx, str(idx))
+            point_items.append({"id": label, "x": float(xy[0]), "y": float(xy[1])})
+        outputs.append({"frame_idx": 0, "points": point_items})
+
+    return outputs
+
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(description="Run basketball court keypoint detection (YOLO pose ONNX).")
+    parser.add_argument(
+        "--stdin-raw",
+        action="store_true",
+        default=True,
+        help="Read raw image bytes from stdin.",
+    )
+    return parser
+
+
+if __name__ == "__main__":
+    build_parser().parse_args()
+
+    base_dir = Path(__file__).resolve().parent
+    model = load_model()
+    if model is None:
+        print("[]")
+        sys.exit(0)
+
+    try:
+        image = load_image(sys.stdin.buffer.read(), base_dir)
+    except Exception:
+        print("[]")
+        sys.exit(0)
+
+    frame = np.array(image)
+    output = run_model(model, frame)
+    print(json.dumps(output))

model_type.json

@@ -0,0 +1,4 @@
+{
+    "project_task_type": "keypoint-detection",
+    "model_type": "yolov11m-pose"
+}

pyproject.toml

@@ -0,0 +1,11 @@
+[project]
+name = "basketball-court-detection-2-19"
+version = "0.1.0"
+requires-python = ">=3.11"
+dependencies = [
+  "numpy>=1.26",
+  "pillow>=10.0",
+  "ultralytics>=8.0.0",
+  "onnx>=1.16",
+  "onnxruntime>=1.17",
+]

weights.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c00051bd207f5b76c6e492f3672be8dd7cce5a217ab271e971378fca057d26d
+size 86282807