In fluorescence-guided surgery, tumors are being made visible to surgeons by using fluorescent near-infrared contrast agents. In the state of the art, these contrast agents are imaged by a normal camera combined with a near-infrared filter.
Fluorescence-lifetime imaging can be used to increase the contrast between a tumor and the background in fluorescence-guided surgery. It is very important for a surgeon that the tumor is clearly visibly so it can be correctly removed without leaving any malignant tissue behind. To image fluorescence lifetime, the fluorescence needs to be excited with a fast pulsed laser and a very special camera that can detect the sub-nanosecond decay behavior of the fluorescence and this with good sensitivity in near infrared. The ETRO department is developing such a camera based on their new CAPS fast-gated image sensor. In theory a CAPS sensor can measure the rate of decay (fluorescence solution decays mono-exponentially after excited by a sharp LASER pulse) by time-gating the fluorescence decay in two time windows and calculating the lifetime from the ratio of the fluorescence emission in the two windows.
During my PhD I investigate how we can develop flexible lifetime estimation methods/algorithms such as CNN deep learning, and machine learning for our fast-gated camera that can use a configurable number of window and averaging to realize a trade-off between frame rate, accuracy and precision. Furthermore, redesign the optical setup to concurrent RGB image, and FLNIR image (intensity, and lifetime), and also the illumination setup.