Research is fun. We take pleasure in creating technology, inventing something useful, and helping people while at the same time learning new skills and teaching them to others. We aim to contribute ideas in support of biomedical imaging, mobile, and remote sensing applications. We specialize on objective and quantitative modeling of data from imaging and other types of sensors by incorporating knowledge from multiple disciplines including applied mathematics, signal processing, machine learning and statistics. These are some of our current research projects:
Lagrangian transforms for signal analysis and machine learning: extracting quantitative and meaning information from signal data is at the heart of many modern problems in science and technology. We are developing new invertible signal transforms, with well defined forward (analysis) and inverse (synthesis) operations. As opposed to linear techniques such as the Wavelet and Fourier transforms, which model a function as a linear combination of fixed basis functions, and thus compares intensities at fixed grids, the new signal transformation allows for quantitative comparison of intensities accross multiple coordinates, and thus are able to capture shifts and other nonlinearities in signal data.
- Project page
- Wasserstein-based time delay estimation, IEEE SPL 2019 Journal
- Generalized sliced Wasserstein distances, preprint, 2019. preprint
- Cumulative Distribution Transform: paper (ACHA 2018)
- Overview paper, IEEE Signal Processing Magazine, 2017. Journal preprint
- Radon Cumulative Distribution Transform: paper (IEEE TIP 2016)
Model-based inverse problems: the Lagrangian signal transformation framework (see project page) offers a powerful mechanism to investigate signal displacement and transport in a mathematically rigorous way. It this enable new modeling approaches that can sumarize image content in very few parameters. We are exploiting these properties do devise new strategies for reconstructing images from limited samples. Applications include estimating facial images, real time cardiac MRI reconstruction, and others, from very few measurements.
- Reconstructing High-Resolution Cardiac MR Movies from Under Sampled Frames, Asilomar, 2017. paper
- Transport-based face reconstruction. IEEE CVPR 2015, pp 4876-4884. pdf
SmokeDetective: Annually in the US, roughly 1 in every 330 households reported a home fire. There is a house fire every 98 seconds and a death once every three hours. Even though over 7 billion in property damage happens every year, 30% of smoke alarms do not work when tested, and smoke alarms are not sounded in 47% of house fires. We are working on algorithms to render any camera enabled smart device (e.g. cell phone, tablet, computer, nanny cam, etc.) into a working smoke detector thus delivering constant, ubiquitous smoke detection and monitoring.
Imaging in turbulence: In collaboration with colleagues at the Naval Research Laboratory, we're developing a new turbulence image formation model which can be used in practical applications. The modeling approach combines stochastic processes with fluid mechanics approaches and is being applied to inverse problems, communications, and target recognition.
- Transport model for beam broadening in turbulence. Modern optics, 2019. paper
- Transport model of atmospheric turbulence, Applied Optics, 2018. paper
- OAM demultipliexing using optimal transport, Optics Express, 2018. paper
Biometrics and stroke : Stroke is a leading cause of disability worldwide. Early recognition and treatment of stroke decreases the mortality and possibility of severe injury. A significant barrier to accurate diagnosis in neurological disease is that visualized deficits such as facial weakness or limb incoordination can be difficult to quantify. Together with collaborators from Neurology at UVa, we are collecting quantitative data from a large patient cohort, as well as controls, and are developing biometric intelligent systems to more objectively quantify signs of stroke and ultimately deliver a low cost, early stroke risk assessment pervasive technology.
- Project page
- Pathological facial weakness detection using computational image analysis, ISBI, 2018. paper
- Robust efficient estimation of heart rate pulse from video. Biomedical Optics Express 2014. paper
Digital pathology, predictive modeling, and high content screening: Digital microscopy is helping revolutionize cancer research, diagnosis, and treatment. Quantitative image analysis methods can help extract objective quantitative information from large quantities of cells. This information can in turn be used to build predictive models to describe effects of perturbations (e.g. drugs), as well as for aiding diagnosis and prognosis from histology/cytology images.
- Are image analysis methods for automated disgnosis in pathology robust? JPI paper
- Nuclear segmentation in pathology images: JPI paper , software
- Epithelium-stroma classification using CNNs. IEEE JBHI, 2017. journal
- Melanoma detection. paper
- Diagnosing mesothelioma from cytology, Cytometry A, 87(4), 326-333, 2015. pdf
- Nuclear morphology in thyroid follicular lesions. Medical Image Analysis, 2014. link
Image-based cytometry and cell modeling: Modern imaging techniques are able to measure information regarding cellular processes with increasing accuracy, and specificity. Numerous applications in health sciences (drug discovery, genetic screens, diagnosis, prognosis, etc.) can be benefited by image data analysis techniques capable of deriving relevant biological information from such datasets. We are developing new approaches for mining information contained in cell image databases, and utilizing it to model important cellular processes.
- Project page: CellOrganizer
- Filament localization and extraction from microscopy images. html
- Automated Learning of Subcellular Variation among Punctate Protein Patterns and a Generative Model of their Relation to Microtubules, PLOS Comp. Biol, 2015. link
- Joint Modeling of Cell and Nuclear Shape Variation, Molecular Biology of the Cell, mbc. E15-06-0370, 2015. Journal site
Transport-based morphometry (TBM): Image-based models of the structure of biological forms can greatly facilitate our understanding of living systems. Based on a new signal transformation being developed (see here for more information) we're developing a new objective approach for quantifying the structure (i.e. shape, texture information) of biological forms seamlessly from image data. The new approach is ideally suited for forms which are associated with transport processes, though it can be used more generally.
- Pre-symptomatic early detection of osteoarthritis from knee MRIs. PNAS, 2020. paper
- Discovery and visualization of structural biomarkers from MRI using transport-based morphometry, Neuroimage 2018. arXiv preprint arXiv:1705.04919, 2018. preprint
- Transport-based morphometry of cells, PNAS 2014. pdf , publisher , software