Facial recognition for disease diagnosis using a deep learning convolutional neural network: a systematic review and...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Computer Vision

Background:

• Deep learning (DL) networks are advancing rapidly, increasing interest in facial recognition technology within healthcare.
• Facial recognition applications are being explored for diagnosing various medical conditions.

Purpose of the Study:

• To systematically review the past decade of literature on DL-based facial recognition for diagnosing rare dysmorphic diseases and facial paralysis.
• To evaluate the effectiveness and applicability of DL facial recognition in disease identification.

Main Methods:

• Systematic literature review following PRISMA guidelines, searching PubMed and other databases.
• Keywords: deep learning convolutional neural networks, facial recognition, disease recognition.
• Screened 208 articles, selected 22 for meta-analysis using Stata 14.0.

Main Results:

• Analyzed 22 studies with 57,953 cases (43,301 diseased samples).
• Meta-analysis revealed a 91.0% accuracy rate (95% CI: 87.0%–95.0%) for DL facial recognition in disease diagnosis.

Conclusions:

• DL-based facial recognition technology demonstrates high accuracy in disease diagnosis.
• Results provide a foundation for further development and clinical application of this technology.