Â鶹´«Ã½ÊÓƵÍøÕ¾

As humans are homeothermic, our internal temperature is closely linked with numerous physiological and psychological mechanisms. Given this, human thermal patterns have been explored to improve the understandings of our body for a couple of centuries (Cho, 2019). Researchers have shown that physiological signatures can be captured through non-contact thermal imaging (e.g. respiration monitoring in Cho et al. 2017). Whilst other contactless sensing methods, such as, RGB camera-based photo-plethysmography suffer from illumination and privacy issues, thermography is much less affected by those constraints (Lloyd, 2013). In addition, studies have shown that different types of physiological activities can be read through the use of a thermal imaging channel (Garbey et al., 2007, Pavlidis et al., 2012, Cho et al. 2017-2019), including respiratory, cardiovascular, perspiratory responses.

In our lab, we have been mainly investigating how to improve reliability, accuracy in physiological measurements through thermal imaging (to support real-world situations). In parallel, we have been building ML frameworks for remote PPG (photo-plethysmography) (with normal RGB cameras eg. webcam, smartphone cameras).

Checkout our recent work:

Self-adversarial Multi-scale Contrastive Learning for Semantic Segmentation of Thermal Facial Images

Low-cost PPG toolkit for Heart Rate and Heart Rate Variability monitoring