Ultrafast Hyperspectral Imaging for Biomedical Optics – ULHYB

In recent decades, there has been a tremendous increase in the spatial resolution of array sensors, which can now exceed 100 megapixels at low cost, and which has revolutionized biomedical imaging. As spatial resolution has become less and less of an issue, the spectral imaging capabilities have remained as a blind spot. Current spectral imaging technologies lack sufficient spectral resolution, speed, robustness to noise, versatility, and cost effectiveness to change the present biomedical paradigm, which is based on few-color imaging.

The main objective of this project is to reshape spectral imaging, by developing a computational framework for high spectral resolution imaging, where physics, computer science, and applied mathematics are put together, to make it possible to image 10 to 1,000 times faster than with current techniques, while also resolving weak signals that cannot be detected at the present time. Motivated by recent advances and deep learning, we propose to shape and manipulate light with spatial light modulators, to acquire multiple pixels at the same time, to provide improved signal-to-noise ratios, while our algorithms will be used to reconstruct high spectral resolution images in real time.

In this project, we will bring this new framework for hyperspectral imaging to the surgical operating room. By providing hyperspectral information in real time during neurosurgery, we will limit the requirement to time-consuming biopsies and help the surgeon to perform safer and less invasive surgery, which will in turn improve survival rates for the patients.