Machine learning for imaging skin cancer with reflectance confocal microscopy


RCM is currently being used in the clinic to assess skin cancer margins as well as provide diagnosis. RCM can optically section a sample which allows the cellular structure to be seen without any contributions from tissue outside of the plane of focus. Thin optical sections require the use of high-NA objectives which are complex, heavy, and expensive. My hope is that, by using AI, thick optical sections collected with small, inexpensive optics, can be sharpened to produce images similar to those obtainable with high numerical aperture.