Using a Pretrained Detection Model

# Installation: pip install ultralytics
from ultralytics import YOLO
import cv2

def detect_with_yolov8(image_path, output_path="yolo_result.jpg",
                        confidence=0.5, iou_threshold=0.45):
    """
    Run YOLOv8 object detection on an image.
    First run: automatically downloads the pretrained model.
    """
    # Load YOLOv8 nano model (fastest, good for learning)
    # Options: yolov8n, yolov8s, yolov8m, yolov8l, yolov8x (nano to xlarge)
    model = YOLO("yolov8n.pt")  # Downloads ~6MB on first run

    # Run detection
    results = model(
        image_path,
        conf=confidence,    # Confidence threshold
        iou=iou_threshold   # IoU threshold for NMS
    )

    print(f"=== YOLOv8 Detection Results ===")
    for result in results:
        boxes = result.boxes
        print(f"Detected {len(boxes)} objects:")

        for box in boxes:
            class_id = int(box.cls[0])
            class_name = model.names[class_id]
            conf = float(box.conf[0])
            x1, y1, x2, y2 = box.xyxy[0].tolist()

            print(f"  {class_name}: {conf:.2%} at ({x1:.0f},{y1:.0f}) to ({x2:.0f},{y2:.0f})")

        # Save annotated image
        annotated = result.plot()  # Returns numpy array with annotations
        cv2.imwrite(output_path, annotated)
        print(f"Saved: {output_path}")

detect_with_yolov8("photo.jpg")