Reproduce “Optimizing YOLOv8 for Parking Space Detection: Comparative Analysis of Custom YOLOv8 Architecture” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Apar Pokhrel, Gia Dao
Researchers from The University of Texas at Arlington systematically evaluated four different backbone architectures—ResNet-18, VGG16, EfficientNetV2, and GhostNet—integrated with YOLOv8 for parking space occupancy detection. Their findings indicate that YOLO-EfficientNetV2 achieved the highest detection accuracy (mAP50:95 of 0.847), while YOLO-GhostNet offered superior computational efficiency, making it suitable for resource-constrained environments.
Key insights
- A distinct trade-off exists between detection accuracy and computational efficiency across the evaluated backbone architectures, indicating no single model is optimal for all scenarios.
- Each backbone demonstrated specific strengths: EfficientNetV2 for highest accuracy, ResNet-18 for a strong balance of accuracy and efficiency, and GhostNet for superior computational efficiency.
- The choice of backbone architecture for real-world parking systems should align with specific deployment requirements, such as available computational resources and real-time processing needs.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.