AI analysis of dorsal hand and facial images predicts chronological age in Indian females with high accuracy, identifying key ageing markers. Hand images, showing knuckle texture as a significant feature, offer a viable alternative to facial images for age prediction, supporting applications in healthspan and longevity research.
February 2024 – Experimental Dermatology
Key takeaways
- Hand images predict age effectively: Dorsal hand images, particularly knuckle texture, can predict chronological age with high accuracy, offering an alternative to facial analysis. This is useful for non-invasive ageing assessments and tracking healthspan, especially in scenarios where facial images are unavailable or impractical, such as in medical or forensic contexts
- Facial features remain key ageing markers: Inner eye corners and mouth areas are crucial for age prediction, highlighting how facial changes reflect biological ageing and can inform personalised wellness strategies. These features, including wrinkles and skin texture, provide valuable insights into the ageing process and potential interventions for maintaining a youthful appearance
- Skin tone impacts age prediction accuracy: Darker skin tones on hands showed higher accuracy in age prediction, suggesting skin pigmentation plays a role in how ageing markers are detected and analysed. This finding underscores the importance of considering diverse skin tones in ageing research to ensure accurate and inclusive healthspan assessments
- AI identifies specific ageing biomarkers: AI analysis revealed distinct hand and facial features, like wrinkles and skin texture, that correlate with ageing, providing insights into visible signs of longevity and healthspan. These biomarkers can help in developing targeted anti-ageing treatments and monitoring the effectiveness of wellness interventions over time
Read the article at: Georgievskaya, Anastasia, et al. “Predicting Human Chronological Age via AI Analysis of Dorsal Hand Versus Facial Images: A Study in a Cohort of Indian Females.” Experimental Dermatology, vol. 33, 2024, e15045. DOI:10.1111/exd.15045.