Combining Deep Learning With Optical Coherence Tomography Imaging to Determine Scalp Hair and Follicle Counts
Hair plays a substantial role in defining one’s identity, representing age, social status, and even wisdom. Effluvium refers to a pathophysiological process that leads to either reversible or permanent hair loss. Not surprisingly, alopecia patients often encounter great amounts of psychological and social morbidity directly related to their illness. These patients’ grief and frustration have, in part, motivated the development of a wide range of hair loss treatment products; however, techniques to determine treatment efficacy are still lacking. Current methods for monitoring hair loss are invasive and or biased in interpretation.
In this study, researchers used optical coherence tomography, or OCT. OCT is a non‐invasive light‐based imaging technique that uses the interference properties of skin tissue and infrared light to achieve high‐resolution cross‐section images that can be evaluated in multiple dimensions, and has the added benefit of not requiring any hair colorization or shaving, allowing for a traceless procedure. The relatively high‐resolution images generated by OCT make it a useful tool for evaluation and diagnosis for many skin conditions.
The purpose of this research was to construct and evaluate a machine‐learning algorithm that could use data obtained from OCT images to accurately count hairs, hair‐bearing follicles, and non‐hair‐bearing follicles. The process automatically returns all counts within seconds of completing the OCT scan of a patch of the scalp and with an accuracy that is within the discrepancy range of human raters.
This research suggests that this approach is well‐positioned to become the standard for non‐invasive evaluation of hair growth treatment progress in patients, saving significant amounts of time and effort compared with manual evaluation.