Coherent Raman hyperspectral imaging technologies have progressed dramatically in recent years, collecting 100’s to 10,000’s of spectra per second with the spectra breadth of traditional spontaneous Raman spectroscopy. There is, however, a lack in available analysis and processing capabilities to bridge the gap between spectroscopy and chemical imaging, in which the end-user is interacting with molecular targets of interest. In this talk we will discuss our latest developments towards this goal, in particular: spectral unmixing/endmember extraction methods and high-speed, high-throughput peak characterization (peak-finding and fitting). Spectral unmixing methods aim to uncover pure species spectra. Certain demonstrated methods, such as vertex component analysis (VCA) require at least 1 pure pixel per chemical is present in the image. Other methods rely on statistical or geometric methods to estimate the pure spectra when no pure pixels are present. In this presentation, we will quantitatively compare results using several state-of-the-art techniques (internally and externally-developed). To autonomously examine retrieved pure spectra, we have developed a high-speed peak finding and fitting algorithm capable of characterizing spectra in micro- to milliseconds, in order to interface with our developed database and data mining methods. Collectively, these developments enable high-speed, high throughput analysis of 1 or many images. Numerical and experimental demonstrations will be presented on an open-source numerical tissue phantom and ~900-color BCARS imagery of murine tissue and clinical specimens.
Charles H. Camp, Sean McIntyre, and Marcus T. Cicerone, "From spectroscopy to chemical imaging: machine learning for hyperspectral coherent Raman imagery (Conference Presentation)," Proc. SPIE 10498, Multiphoton Microscopy in the Biomedical Sciences XVIII, 104981E (Presented at SPIE BiOS: January 30, 2018; Published: 14 March 2018); https://doi.org/10.1117/12.2291080.5751624476001.
Conference Presentations are recordings of oral presentations given at SPIE conferences and published as part of the conference proceedings. They include the speaker's narration along with a video recording of the presentation slides and animations. Many conference presentations also include full-text papers. Search and browse our growing collection of more than 14,000 conference presentations, including many plenary and keynote presentations.
Study of self-shadowing effect as a simple means to realize nanostructured thin films and layers with special attentions to birefringent obliquely deposited thin films and photo-luminescent porous silicon