AI bot to detect fake COVID-19 vaccine certificate

Area of Science:

• Computer Science
• Artificial Intelligence
• Biotechnology

Background:

• The widespread need for COVID-19 vaccination certificates highlights a vulnerability to forgery.
• Fake vaccine certificates pose significant public health risks and strain healthcare systems.
• Distinguishing authentic from fraudulent vaccination proof is crucial for pandemic management.

Purpose of the Study:

• To develop an Artificial Intelligence (AI) and Deep Learning-based system for detecting forged COVID-19 vaccine certificates.
• To create a reliable method for verifying the authenticity of digital vaccine certificates.
• To mitigate the risks associated with the circulation of fake vaccination proof.

Main Methods:

• A multi-stage approach involving data collection, preprocessing, error level analysis, and texture-based feature extraction (logo, symbol, signature Crest-Trough parameter).
• Classification of certificates as real or fake using the DenseNet201 deep learning architecture.
• Evaluation against state-of-the-art models including SVM, RNN, VGG16, Alexnet, and CNN.

Main Results:

• The proposed DenseNet201 and Local Binary Patterns (D201-LBP) model achieved a high accuracy of 0.94.
• The model demonstrated superior performance across various metrics such as specificity, sensitivity, detection rate, recall, and f1-score compared to existing methods.
• Efficient computation time was observed in the proposed model.

Conclusions:

• The AI-powered system effectively detects fake vaccine certificates with high accuracy.
• This technology offers a robust solution to combat vaccine certificate fraud.
• The findings support the implementation of advanced digital verification systems for public health documentation.