I found more details at https://www.rouast.com/blog/articles/vitallens-take-a-vital-selfie/ https://www.rouast.com/blog/articles/vitallens-take-a-vital-... and

I found more details at https://www.rouast.com/blog/articles/vitallens-take-a-vital-selfie/ https://www.rouast.com/blog/articles/vitallens-take-a-vital-... and https://arxiv.org/pdf/2312.06892 https://arxiv.org/pdf/2312.06892. It's still vague beyond saying they trained an AI: "VitalLens is trained on the PROSIT and Vital Videos Africa dataset...The time-synchronized sensor array used for PROSIT consists of a video camera, electrocardiogram (ECG), pulse oximetry, blood pressure monitor, and an ambient light sensor. "