A study on the recognition of monkeypox infection based on deep convolutional neural networks

Area of Science:

• Artificial Intelligence
• Medical Imaging
• Epidemiology

Background:

• The World Health Organization (WHO) reported over 14,000 monkeypox cases in 71 countries, with global risk assessed as moderate.
• Current monkeypox diagnosis relies on traditional methods, facing challenges in efficiency and cost.
• Deep learning excels at rapid and reliable image feature extraction and recognition.

Purpose of the Study:

• To develop a novel deep learning model for accurate monkeypox image recognition.
• To address the limitations of traditional diagnostic methods for monkeypox.

Main Methods:

• A residual convolutional neural network incorporating a λ function and contextual transformer (LaCTResNet) was proposed.
• The model was evaluated for its image recognition capabilities on monkeypox cases.

Main Results:

• The LaCTResNet model achieved an average recognition accuracy of 91.85%.
• This accuracy represents a 15.82% improvement over the baseline ResNet50 model.
• Performance surpassed other classical convolutional neural networks, including AlexNet, VGG16, Inception-V3, and EfficientNet-B5.

Conclusions:

• The developed LaCTResNet method enables high-precision identification of monkeypox skin symptoms.
• This provides a fast, reliable auxiliary diagnostic method for front-line medical staff.
• The study highlights the potential of AI in enhancing infectious disease diagnosis.