The amount of information that can be extracted from the images captured by our setups can be maximized by harnessing the computational techniques on the images post-acquisition. At the Biophotonics Imaging Laboratory, we utilize both mathematical models of image formation and advances in artificial intelligence and machine learning to not only enhance the quality of images captured but also to automate the translation of these images to meaningful biological information.
Previously, we have devised solutions to the scattering problem in optical coherence tomography (OCT) to computationally restore the effect of depth-of-focus for entire OCT volumes hand have used it for tumor imaging, cellular imaging, and ophthalmology. Additionally, we have used concepts of wavefront engineering to not only correct optical aberrations computationally, but to automate wavefront sensing from OCT images directly and use aberrations as a novel contrast mechanism in OCT. We have utilized these not just for OCT, but also for non-linear and multiphoton imaging modalities.
In recent years, the burgeoning field of machine learning and artificial intelligence have provided researches with tools that can discern the pattern in our images to recognize the relevant information much easier than rigorous mathematical models. We have utilized these techniques for numerous applications for automating the diagnosis of cancer for autofluorescence multi-harmonic multiphoton microscopy, and for polarization-sensitive OCT of breast tumors. Additionally, we have applied pattern recognition and machine learning models for the detection of middle ear effusions and for metabolic imaging.
Finally, the setups that we design at the Biophotonics Imaging Laboratory generate highly multidimensional data. For some of our setups, the raw data is generated at thousands of megabytes per second. Therefore, building data analysis pipelines and real-time processing techniques to deal with the data generated efficiently is a critical aspect of our research.
Aberrations are the wavefront phase deviations of the light from the desired ideal shape that cause imperfect image formation in optical microscopes. They are caused either by imperfections in optics in the imaging systems or by the sample structure. Typically, in the context of biomedical imaging, sample induced aberrations limit the system performance by reducing the maximum imaging depth, image sharpness and contrast. Read more...
In BIL, we have been generating multi-channel images using our unique optical imaging systems. Considering the rich content our images have, it is challenging for a human to examine every layer of those images and extract all the useful information that is potentially relevant to the question we are investigating. With the assistance of Artificial Intelligence (AI), we are now able to accomplish those tasks by training and implementing different types of Machine Learning (ML) and Deep Learning (DL) models. Read more...