Spaces:
Running
on
Zero
Running
on
Zero
File size: 2,559 Bytes
b23251f |
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 |
import os
import numpy as np
from PIL import Image, ImageDraw
import logging
from ultralytics import YOLO
from utils.model_downloader import download_model_if_needed
logger = logging.getLogger(__name__)
class ObjectDetector:
"""
Generalized Object Detection Wrapper for YOLOv5, YOLOv8, and future variants.
"""
def __init__(self, model_key="yolov5n-seg", weights_dir="models/detection/weights", device="cpu"):
"""
Initialize the Object Detection model.
Args:
model_key (str): Model identifier as defined in model_downloader.py.
weights_dir (str): Directory to store/download model weights.
device (str): Inference device ("cpu" or "cuda").
"""
weights_path = os.path.join(weights_dir, f"{model_key}.pt")
download_model_if_needed(model_key, weights_path)
logger.info(f"Loading Object Detection model '{model_key}' from {weights_path}")
self.device = device
self.model = YOLO(weights_path)
def predict(self, image: Image.Image):
"""
Run object detection.
Args:
image (PIL.Image.Image): Input image.
Returns:
List[Dict]: List of detected objects with class name, confidence, and bbox.
"""
logger.info("Running object detection")
results = self.model(image)
detections = []
for r in results:
for box in r.boxes:
detections.append({
"class_name": r.names[int(box.cls)],
"confidence": float(box.conf),
"bbox": box.xyxy[0].tolist()
})
logger.info(f"Detected {len(detections)} objects")
return detections
def draw(self, image: Image.Image, detections, alpha=0.5):
"""
Draw bounding boxes on image.
Args:
image (PIL.Image.Image): Input image.
detections (List[Dict]): Detection results.
alpha (float): Blend strength.
Returns:
PIL.Image.Image: Image with bounding boxes drawn.
"""
overlay = image.copy()
draw = ImageDraw.Draw(overlay)
for det in detections:
bbox = det["bbox"]
label = f'{det["class_name"]} {det["confidence"]:.2f}'
draw.rectangle(bbox, outline="red", width=2)
draw.text((bbox[0], bbox[1]), label, fill="red")
return Image.blend(image, overlay, alpha)
|