Update yolov8.py
Browse files
yolov8.py
CHANGED
@@ -0,0 +1,153 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import cv2
|
3 |
+
import numpy as np
|
4 |
+
import supervision as sv
|
5 |
+
from ultralytics import YOLO
|
6 |
+
import yaml
|
7 |
+
from pathlib import Path
|
8 |
+
import torch
|
9 |
+
|
10 |
+
print(torch.cuda.is_available())
|
11 |
+
|
12 |
+
def setup_dataset_config(dataset_path, class_names):
|
13 |
+
data_yaml = {
|
14 |
+
'path': os.path.abspath(dataset_path),
|
15 |
+
'train': 'train/images',
|
16 |
+
'val': 'valid/images',
|
17 |
+
'test': 'test/images',
|
18 |
+
'names': {i: name for i, name in enumerate(class_names)},
|
19 |
+
'nc': len(class_names)
|
20 |
+
}
|
21 |
+
|
22 |
+
with open(os.path.join(dataset_path, 'dataset.yaml'), 'w') as f:
|
23 |
+
yaml.dump(data_yaml, f, sort_keys=False)
|
24 |
+
|
25 |
+
print(f"Dataset config saved to {os.path.join(dataset_path, 'dataset.yaml')}")
|
26 |
+
return os.path.join(dataset_path, 'dataset.yaml')
|
27 |
+
|
28 |
+
|
29 |
+
def train_yolov8_model(dataset_config, epochs=100, img_size=640, batch_size=16):
|
30 |
+
model = YOLO('yolov8n.pt')
|
31 |
+
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
32 |
+
print(f"Training on device: {device}")
|
33 |
+
|
34 |
+
results = model.train(
|
35 |
+
data=dataset_config,
|
36 |
+
epochs=epochs,
|
37 |
+
imgsz=img_size,
|
38 |
+
batch=batch_size,
|
39 |
+
name='accessory_detection',
|
40 |
+
patience=20,
|
41 |
+
save=True,
|
42 |
+
device=device,
|
43 |
+
verbose=True
|
44 |
+
)
|
45 |
+
|
46 |
+
print("Training completed!")
|
47 |
+
return model
|
48 |
+
|
49 |
+
|
50 |
+
def validate_model(model):
|
51 |
+
metrics = model.val()
|
52 |
+
print(f"Validation metrics: {metrics}")
|
53 |
+
return metrics
|
54 |
+
|
55 |
+
|
56 |
+
def run_webcam_detection(model_path=None):
|
57 |
+
if model_path is None:
|
58 |
+
runs_dir = Path('runs/detect')
|
59 |
+
if runs_dir.exists():
|
60 |
+
model_dirs = [d for d in runs_dir.iterdir() if d.is_dir() and d.name.startswith('accessory_detection')]
|
61 |
+
if model_dirs:
|
62 |
+
latest_model = max(model_dirs, key=os.path.getmtime) / 'weights' / 'best.pt'
|
63 |
+
if latest_model.exists():
|
64 |
+
model_path = str(latest_model)
|
65 |
+
print(f"Using latest model: {model_path}")
|
66 |
+
|
67 |
+
model = YOLO(model_path) if model_path else YOLO('yolov8n.pt')
|
68 |
+
print(f"Model loaded from {model_path if model_path else 'Pretrained YOLOv8n'}")
|
69 |
+
|
70 |
+
cap = cv2.VideoCapture(0, cv2.CAP_V4L2)
|
71 |
+
if not cap.isOpened():
|
72 |
+
print("Error: Could not open webcam.")
|
73 |
+
return
|
74 |
+
|
75 |
+
box_annotator = sv.BoxAnnotator(thickness=2, text_thickness=2, text_scale=1)
|
76 |
+
print("Press 'q' to quit")
|
77 |
+
|
78 |
+
while True:
|
79 |
+
ret, frame = cap.read()
|
80 |
+
if not ret:
|
81 |
+
print("Error: Failed to capture image")
|
82 |
+
break
|
83 |
+
|
84 |
+
results = model(frame, conf=0.25)
|
85 |
+
detections = sv.Detections.from_ultralytics(results[0])
|
86 |
+
class_names = model.names if hasattr(model, 'names') else {0: "unknown"}
|
87 |
+
|
88 |
+
labels = [
|
89 |
+
f"{class_names[class_id]} {confidence:.2f}"
|
90 |
+
for _, confidence, class_id, _ in detections
|
91 |
+
]
|
92 |
+
|
93 |
+
frame = box_annotator.annotate(scene=frame, detections=detections, labels=labels)
|
94 |
+
cv2.putText(frame, "Press 'q' to quit", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 2)
|
95 |
+
|
96 |
+
cv2.imshow("YOLOv8 Accessory Detection", frame)
|
97 |
+
|
98 |
+
if cv2.waitKey(1) & 0xFF == ord('q'):
|
99 |
+
break
|
100 |
+
|
101 |
+
cap.release()
|
102 |
+
cv2.destroyAllWindows()
|
103 |
+
|
104 |
+
|
105 |
+
def prepare_custom_dataset(source_dir, target_dir, split_ratios=(0.7, 0.2, 0.1)):
|
106 |
+
import shutil
|
107 |
+
from sklearn.model_selection import train_test_split
|
108 |
+
|
109 |
+
os.makedirs(os.path.join(target_dir, 'train', 'images'), exist_ok=True)
|
110 |
+
os.makedirs(os.path.join(target_dir, 'train', 'labels'), exist_ok=True)
|
111 |
+
os.makedirs(os.path.join(target_dir, 'valid', 'images'), exist_ok=True)
|
112 |
+
os.makedirs(os.path.join(target_dir, 'valid', 'labels'), exist_ok=True)
|
113 |
+
os.makedirs(os.path.join(target_dir, 'test', 'images'), exist_ok=True)
|
114 |
+
os.makedirs(os.path.join(target_dir, 'test', 'labels'), exist_ok=True)
|
115 |
+
|
116 |
+
print("YOLOv8 directory structure created")
|
117 |
+
|
118 |
+
files = [f for f in os.listdir(source_dir) if f.endswith('.txt') and not f.endswith('classes.txt')]
|
119 |
+
|
120 |
+
train_files, temp_files = train_test_split(files, test_size=(split_ratios[1]+split_ratios[2]), random_state=42)
|
121 |
+
val_ratio = split_ratios[1] / (split_ratios[1] + split_ratios[2])
|
122 |
+
val_files, test_files = train_test_split(temp_files, test_size=(1-val_ratio), random_state=42)
|
123 |
+
|
124 |
+
print(f"Split dataset: {len(train_files)} train, {len(val_files)} validation, {len(test_files)} test images")
|
125 |
+
|
126 |
+
setup_dataset_config(target_dir, ["hat", "scarf", "sunglasses", "spectacles", "headphones", "ears_visible"])
|
127 |
+
print("Dataset preparation completed!")
|
128 |
+
|
129 |
+
return os.path.join(target_dir, 'dataset.yaml')
|
130 |
+
|
131 |
+
|
132 |
+
if __name__ == "__main__":
|
133 |
+
import argparse
|
134 |
+
parser = argparse.ArgumentParser(description="YOLOv8 Face Accessory Detection System")
|
135 |
+
parser.add_argument('--train', action='store_true', help='Train model')
|
136 |
+
parser.add_argument('--detect', action='store_true', help='Run detection on webcam')
|
137 |
+
parser.add_argument('--config', type=str, help='Path to dataset config file')
|
138 |
+
parser.add_argument('--model', type=str, help='Path to trained model')
|
139 |
+
parser.add_argument('--epochs', type=int, default=100, help='Number of training epochs')
|
140 |
+
args = parser.parse_args()
|
141 |
+
|
142 |
+
if args.train:
|
143 |
+
if not args.config:
|
144 |
+
print("Error: Dataset config is required for training")
|
145 |
+
else:
|
146 |
+
model = train_yolov8_model(args.config, epochs=args.epochs)
|
147 |
+
validate_model(model)
|
148 |
+
|
149 |
+
if args.detect:
|
150 |
+
run_webcam_detection(args.model)
|
151 |
+
|
152 |
+
if not (args.train or args.detect):
|
153 |
+
parser.print_help()
|