Research

Medical Image Computing and Computer Assisted Intervention

Skin2Mind: A Multimodal Framework for Mental Health Risk Screening via Facial Images and Structured Skin Reports

Mental health screening relies on subjective questionnaires, leaving objective biological signals uncaptured. Grounded in the "skin-brain axis," we present Skin2Mind, the first psychodermatology-based multimodal AI framework for individual mental health risk prediction.

Scientific Reports

Facial dermal exposure-related parameters and skincare behaviors among adult women from six cities in China

This study aimed to describe facial dermal exposure-related parameters and skincare behaviors among adult women in China and to explore their associations with sensitive skin status and objective skin biophysical measurements.

Society for Investigative Dermatology

Prediction of skin biophysical parameters from facial images using deep learning in large-scale population cohorts

Using ~6,000 participants across three independent cohorts, we developed and externally validated machine learning models to predict multidimensional skin phenotypes from facial images.

Journal of Investigative Dermatology

GWASs of the Nasolabial Fold Identified Variants Related to Genes that Also Affect Facial Morphology

The nasolabial fold (NLF) is a prominent dermatological phenotype of the aging midface. Previous anatomical studies have clarified that the NLF is potentially induced by the aging changes in the superficial musculoaponeurotic system architecture, cutaneous ligament, midface musculature and fat compartments, and craniofacial skeleton.

Society for Investigative Dermatology

Evaluation of imaging-based methods for facial aging detection

Quantifying facial aging is essential in dermatology for studying age-related changes and assessing the effectiveness of skincare products. However, facial aging detection remains challenging due to the absence of standardized benchmarks and unified testing protocols, making it difficult to compare different methods fairly.

International Federation of Societies of Cosmetic Chemists

Deep Learning Analysis of Perceived Facial Aging and Influential Features Across Evaluator Groups

Facial aging has long been recognized as an important biomarker of aging and overall health. It is closely linked to numerous age-related diseases, including cardiovascular conditions, metabolic disorders, and neurodegenerative syndromes. As a visible and socially relevant indicator, facial aging has received sustained attention from both scientific communities and the public.

Journal of Cosmetic Dermatology

Transdermal Delivery of Baicalin Based on Bio-Vesicles and Its Efficacy in Antiaging of the Skin

To develop a stable and efficient delivery system for baicalin, a flavonoid with potential antioxidant and antiaging properties, to overcome its limitations in solubility, stability, and skin permeability.

Society for Investigative Dermatology

Identifying Facial Regions and Aging Features Associated with Perceived Age: A Deep Learning-Based Facial Aging Assessment Method (2024)

Facial aging features manifest with considerable inter-individual variability, leading some individuals to appear younger while others appear older. Classic experiments on perceived age rely on human assessment, which demands significant human resources. In this study, we assembled 160 evaluators to assess the perceived age of 3,186 subjects' faces.

International Society for Investigative Dermatology

Identification the inflection points of wrinkle types in a large-scale population study of 431,321 subjects

Using AI-based skin analysis on a dataset of 431,321 subjects, this study identifies critical age inflection points at which different wrinkle types begin to accelerate. Our large-scale population analysis provides actionable benchmarks for age-specific skincare product development and clinical intervention timing.

Research Square

Integration of skin phenome and microbiome reveals the key role of bacteria in human skin aging