A comprehensive dataset for automotive brand/logo detection and recognition, formatted for YOLO training. This dataset contains high-quality images with precise annotations for 33 different automotive brands.

📊 Dataset Overview

This dataset is specifically designed for object detection tasks focusing on automotive brand recognition in various contexts including vehicles, logos, and brand emblems.

📈 Dataset Statistics

Total Images: 1,404
Total Annotations: 1,452
Number of Classes: 33 automotive brands
Image Format: JPG
Annotation Format: YOLO v8 (.txt files)
Split Ratio: Train 75.4% | Valid 15.2% | Test 9.3%

📂 Directory Structure

brand-eye-dataset/
├── train/
│   ├── images/          # 1,059 training images
│   └── labels/          # 1,059 training annotations (.txt)
├── valid/
│   ├── images/          # 214 validation images  
│   └── labels/          # 214 validation annotations (.txt)
├── test/
│   ├── images/          # 131 test images
│   └── labels/          # 131 test annotations (.txt)
├── data.yaml           # Dataset configuration
└── README.md           # This file

🚗 Automotive Brands (Classes)

The dataset includes the following 33 automotive brands:

0: audi          1: bmw           2: byd           3: cherry
4: chevrolet     5: citroen       6: cupra         7: dacia
8: fiat          9: ford          10: honda        11: hyundai
12: kia          13: landrover    14: mazda        15: mercedes
16: mg           17: mini         18: mitsubishi   19: nissan
20: opel         21: pegout       22: porsche      23: rangerover
24: renault      25: seat         26: skoda        27: suzuki
28: tesla        29: togg         30: toyota       31: volvo
32: wolksvogen

📋 YOLO Annotation Format

Each image has a corresponding .txt file with the same name containing bounding box annotations:

class_id center_x center_y width height

Where:

class_id: Integer class identifier (0-32)
center_x, center_y: Normalized center coordinates (0.0-1.0)
width, height: Normalized width and height (0.0-1.0)

Example annotation:

15 0.5 0.3 0.2 0.4    # Mercedes logo at center-left
26 0.7 0.8 0.15 0.25  # Skoda emblem at bottom-right

🚀 Quick Start

Loading with Ultralytics

from ultralytics import YOLO

# Train a model
model = YOLO('yolov8n.pt')
results = model.train(
    data='data.yaml',
    epochs=100,
    imgsz=640,
    batch=16
)

Loading with Custom Code

import os
import cv2
import yaml

def load_dataset_info(data_yaml_path):
    with open(data_yaml_path, 'r') as f:
        data = yaml.safe_load(f)
    return data

def load_image_and_labels(img_path, label_path):
    # Load image
    image = cv2.imread(img_path)

    # Load labels
    labels = []
    if os.path.exists(label_path):
        with open(label_path, 'r') as f:
            for line in f:
                labels.append(list(map(float, line.strip().split())))

    return image, labels

# Example usage
data_info = load_dataset_info('data.yaml')
print(f"Classes: {data_info['names']}")

Visualization Example

import matplotlib.pyplot as plt
import matplotlib.patches as patches
import cv2

def visualize_annotations(img_path, label_path, class_names):
    image = cv2.imread(img_path)
    image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
    h, w = image.shape[:2]

    fig, ax = plt.subplots(1, 1, figsize=(12, 8))
    ax.imshow(image)

    # Read annotations
    with open(label_path, 'r') as f:
        for line in f:
            class_id, cx, cy, width, height = map(float, line.strip().split())

            # Convert to pixel coordinates
            x = (cx - width/2) * w
            y = (cy - height/2) * h
            w_box = width * w
            h_box = height * h

            # Create rectangle
            rect = patches.Rectangle((x, y), w_box, h_box, 
                                   linewidth=2, edgecolor='red', facecolor='none')
            ax.add_patch(rect)

            # Add label
            brand_name = class_names[int(class_id)]
            ax.text(x, y-10, brand_name, 
                   color='red', fontsize=10, fontweight='bold',
                   bbox=dict(boxstyle="round,pad=0.3", facecolor="white", alpha=0.7))

    plt.axis('off')
    plt.tight_layout()
    plt.show()

# Example usage
class_names = ['audi', 'bmw', 'byd', 'cherry', 'chevrolet', 'citroen', 'cupra', 'dacia',
               'fiat', 'ford', 'honda', 'hyundai', 'kia', 'landrover', 'mazda', 'mercedes',
               'mg', 'mini', 'mitsubishi', 'nissan', 'opel', 'pegout', 'porsche', 'rangerover',
               'renault', 'seat', 'skoda', 'suzuki', 'tesla', 'togg', 'toyota', 'volvo', 'wolksvogen']

# visualize_annotations('train/images/sample.jpg', 'train/labels/sample.txt', class_names)

📊 Dataset Quality & Statistics

Split Distribution

Training Set: 1,059 images (75.4%)
Validation Set: 214 images (15.2%)
Test Set: 131 images (9.3%)

Class Distribution

Balanced representation across all 33 automotive brands with varying annotation density based on brand prevalence in real-world scenarios.

🎯 Use Cases

Automotive Brand Recognition: Identify car brands from images
Dealership Management: Automated inventory cataloging
Market Research: Brand presence analysis in automotive industry
Traffic Analysis: Vehicle brand distribution studies
Insurance Applications: Automated vehicle brand identification
Parking Management: Brand-specific parking solutions

🏆 Model Training Results

This dataset was used to train the Brand Eye YOLO model with excellent results:

[email protected]: 0.808 (Final epoch)
[email protected]:0.95: 0.473 (Final epoch)
Training Epochs: 50
Best Performance: Available in model repository
Model Size: ~14MB (optimized for deployment)

🔧 Data Preparation & Validation

Quality Assurance

def validate_dataset(images_dir, labels_dir):
    issues = []

    for img_file in os.listdir(images_dir):
        if img_file.lower().endswith(('.jpg', '.jpeg', '.png')):
            label_file = img_file.rsplit('.', 1)[0] + '.txt'
            label_path = os.path.join(labels_dir, label_file)

            if not os.path.exists(label_path):
                issues.append(f"Missing label: {label_file}")
            else:
                # Check label format
                with open(label_path, 'r') as f:
                    for i, line in enumerate(f):
                        parts = line.strip().split()
                        if len(parts) != 5:
                            issues.append(f"Invalid format in {label_file}:{i+1}")
                        else:
                            try:
                                cls_id, cx, cy, w, h = map(float, parts)
                                if not (0 <= cx <= 1 and 0 <= cy <= 1 and 0 <= w <= 1 and 0 <= h <= 1):
                                    issues.append(f"Invalid coordinates in {label_file}:{i+1}")
                                if not (0 <= cls_id <= 32):
                                    issues.append(f"Invalid class_id in {label_file}:{i+1}")
                            except ValueError:
                                issues.append(f"Non-numeric values in {label_file}:{i+1}")

    return issues

# Validate dataset
train_issues = validate_dataset('train/images/', 'train/labels/')
print(f"Training set validation: {len(train_issues)} issues found")

📄 License

This dataset is released under Apache 2.0 License.

🤝 Citation

If you use this dataset in your research, please cite:

@dataset{brand_eye_dataset_2024,
  title={Brand Eye Dataset: Automotive Brand Detection Dataset},
  author={Haydar Kadioğlu},
  year={2024},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/haydarkadioglu/brand-eye-dataset}
}

🔗 Related Resources

🤖 Trained Model: haydarkadioglu/brand-eye-yolo
📊 Training Analysis: See visualize.ipynb in model repository
💻 Code Examples: Available in model repository
🐛 Issues: Report issues in model repository

📞 Contact

GitHub: haydarkadioglu
Hugging Face: haydarkadioglu
Model Repository: brand-eye-yolo

🎯 Perfect for automotive industry applications, research, and brand recognition tasks!

This dataset represents months of careful curation and annotation work, providing high-quality training data for automotive brand detection systems.

Downloads last month: 25

Models trained or fine-tuned on haydarkadioglu/brand-eye-dataset

haydarkadioglu/brand-eye

Object Detection • Updated Sep 23 • 11 • 3