Convolutional Block Attention Module with Faster R Convolutional Neural Network for Face Recognition in Car Parking System

A face recognition based multi-location Internet of Things (IoT) car parking system is established to enhance parking management across different locations through facial recognition for access control. The IoT system integrates devices to monitor parking spaces and track availability, while facial recognition confirms that only authorized users can park in designated areas, eliminating the need for cards or tickets. This system ensures entry and exit enhancing both user convenience and security, administrators to manage and analyze parking usage efficiently. Then the system produces alerts or sound for unauthorized access and supports dynamic allocation parking of spaces relies on demand. However, ensuring security and accessibility in multi-location parking service is challenging due to risks of unauthorized access and ineffective space utilization. Unauthorized access remains a risk particularly if facial system is enough to prevent or identify fake detection. These challenges are needed enhanced system integration, authentication. To overcome these challenges, Convolutional Block Attention Module- Faster R Convolutional Neural Network (CBAM-FRCNN) is proposed for facial recognition in the car parking system based on IoT. The proposed CBAM-FRCNN improves facial recognition in car parking systems through feature extraction and attention mechanism for enhanced accuracy in functioning. The CBAM helps to focused on necessary spatial and channel wise features in the data which enhancing relevant facial data are extracted. This combination enhances recognition accuracy while ensuring authorized users are allowed for access and maintaining reliable system performance. The proposed method, CBAM-FRCNN achieves an accuracy of 98.5% and an F1-score of 98.6%, thereby outperforming the existing Deep Convolutional Neural Network (DCNN).