The project aims to develop a novel rapid spectral imaging microscopy technique and device in light of its expected unprecedented capability in cancer detection. It brings together the expertise of three research groups at two universities with years of experience in relevant fields. While common image acquisition provides only three values (RGB) for each pixel of the image, spectral imaging provides the full spectrum to every pixel, which is a three-dimensional (3D) data (2D spatial + spectral). It carries valuable information that is critical for many applications. One important example that we intend to pursue is cancer diagnostics of stained biopsies. This application addresses an unmet need in order to cope with the high demand for biopsies diagnostics. The additional spectral information is significant for detecting cancerous cells due to the complexity of the images, the variability of the slides and the unstable imaging conditions.
Currently, the acquisition of microscope spectral images takes long time and it is not even considered for practical use, even though it is believed to be the solution for pathological biopsy diagnostics. In this project a novel spectral microscopy approach will be implemented following the design guidelines of compressive sensing (CS) theory. CS is a revolutionary sampling theorem introduced 12 years ago and it breaks the traditional Nyquist -Shannon sampling paradigm that ruled during the 20th century. Owing to the CS design, the acquisition time of the spectral images with the new technology to be developed is reduced by an order of magnitude.
We will use a commercial brightfield transmission microscope. The new method requires to modify the illumination path of the sample. We will design an illumination unit that can generate different spectral patterns (linear combinations of spectral bands). During the acquisition, multiple images of the specimen are taken, each with a different spectral illumination pattern. The set of the illumination patterns are designed according to the compressing sensing theory; thus the number of exposures is much smaller than the number of spectral bands that are ultimately retrieved by an appropriate reconstruction algorithm.
The key hardware component in the new system is a Liquid Crystal (LC) device that will be implemented in the illumination path of the microscope. Since only the illumination arm of the microscope is changed, the technology is versatile and relatively inexpensive, furthermore, such a spectral modulator can then be used with any other imaging system.
In the framework of the project the spectral illumination will be designed, a new special LC spectral modulator will be investigated and manufactured (IA group), appropriate algorithms will be developed and implemented in a computerised control unit (AS group). Ultimately, after integrating the system, its performance for histopathology will be tested (YG group).