A novel hybrid face mask detection approach using Transformer and convolutional neural network models

Area of Science:

• Computer Vision
• Deep Learning
• Artificial Intelligence

Background:

• Face mask detection is crucial for public safety, with current models often based on convolutional neural networks (CNNs).
• CNNs struggle with low-quality images and capturing long-range dependencies, limiting their effectiveness.
• Transformer neural networks offer enhanced capabilities for handling long-range dependencies in visual data.

Purpose of the Study:

• To develop an automatic hybrid face mask detection model integrating transformer and convolutional neural network architectures.
• To evaluate the performance of the proposed hybrid model against existing state-of-the-art methods.
• To assess the model's suitability for real-time public health applications.

Main Methods:

• Developed a hybrid deep learning model combining transformer and CNN components.
• Trained and evaluated the model on face mask detection tasks.
• Compared the hybrid model's performance metrics (precision, execution time) with other leading models.

Main Results:

• The proposed hybrid model achieved a highest average precision of 89.4%.
• The model demonstrated an efficient execution time of 2.8 seconds.
• Experimental results confirmed the hybrid model's superior performance compared to existing methods.

Conclusions:

• The hybrid transformer-CNN model is effective for accurate face mask detection.
• The model's high precision and speed make it suitable for practical, real-time applications.
• This technology can significantly contribute to infectious disease control and public healthcare efforts.