Screening of Genetic Syndromes by Facial Recognition Technology: VGG-16 Screening Model Construction and Evaluation

Abstract

Background: Many genetic syndromes (GSs) have distinct facial dysmorphism, and facial gestalts can be used as a diagnostic tool for recognizing a syndrome. Facial recognition technology has advanced in recent years, and the screening of GSs by facial recognition technology has become feasible. This study constructed a facial recognition model for GSs screening. Methods: One hundred and fifty-four children with GSs and 169 healthy children were enrolled in this study. A total of 506 frontal facial photos were collected and divided into training and testing datasets. The VGG-16 network was pre-trained by the VGG-Face dataset. Model parameters were fine-tuned with the training dataset, and a facial recognition model based on the VGG-16 architecture was constructed. The performance of the VGG-16 model was evaluated, and tested by pediatricians. Findings: The VGG-16 model achieved the highest accuracy of 0•92, specificity 0•914, sensitivity 0•923, F1-score 90•4%, and an area under the receiver operating characteristic curve of 0•92 for GSs screening, which was significantly higher than that achieved by human experts. Interpretation: This study highlighted the feasibility of facial recognition technology for GSs identification. The VGG-16 recognition model could play a prominent role in GSs screening in clinical practice. Funding Statement: This study was supported by the National Natural Science Foundation of China (Grant No.82070321), Sanming Project of Medicine in Shenzhen (CN) (Grant No.H022017031). Declaration of Interests: All the authors declare that they have no conflicts of interests. Ethics Approval Statement: This study was approved by the Research Ethics Committee of Guangdong Provincial Peoples’ Hospital (Project Number: KY2020-033-01). Informed consent was given by all patients or their wardens to analyze and publish the pictures.