Optimizing YOLOv8 for Parking Space Detection: Comparative Analysis of Custom YOLOv8 Architecture

TL;DR

This paper compares different custom backbone architectures integrated with YOLOv8 to improve parking space occupancy detection, analyzing their accuracy and efficiency for better parking management systems.

Contribution

It provides a comprehensive evaluation of various backbone architectures with YOLOv8 for parking detection, highlighting their strengths and trade-offs.

Findings

ResNet-18 offers high accuracy with moderate computation.

EfficientNetV2 achieves a good balance between speed and detection performance.

Ghost backbone provides faster inference with slightly lower accuracy.

Abstract

Parking space occupancy detection is a critical component in the development of intelligent parking management systems. Traditional object detection approaches, such as YOLOv8, provide fast and accurate vehicle detection across parking lots but can struggle with borderline cases, such as partially visible vehicles, small vehicles (e.g., motorcycles), and poor lighting conditions. In this work, we perform a comprehensive comparative analysis of customized backbone architectures integrated with YOLOv8. Specifically, we evaluate various backbones -- ResNet-18, VGG16, EfficientNetV2, Ghost -- on the PKLot dataset in terms of detection accuracy and computational efficiency. Experimental results highlight each architecture's strengths and trade-offs, providing insight into selecting suitable models for parking occupancy.