Tristan Swedish

Computational Model of Optical Scattering by Elastin in Lung

Little is understood about the detailed micromechanical properties of lung in vivo. Attempts to improve imaging are hampered by heterogeneity of the tissue. One common ex vivo technique is optical coherence tomography (OCT). Simulated OCT with a Finite-Difference Time-Domain (FDTD) computer model elucidates the relationship between captured images and the physical geometry of the lung. Parallel computation and improved processing power make accurate coherent imaging models feasible. A previous FDTD model of pulsed laser wave propagation in the lung produced images that displayed many of the properties of experimental images. The model was improved with the addition of elastin and increased computational volume. Elastin plays an important role in the simulation because the combination of its fibrous structure and high index of refraction acts as an excellent scatterer of light. This strong scattering increases the signal reported by the simulated OCT scan in areas where elastin is most abundant, improving visualization of the structure as more light is reflected back from the heterogeneous elastin network. However, scattering by elastin decreases the depth of penetration and leads to images that are more difficult to interpret. Gaining a better understanding of how lung structures affect light propagation will lead to improved signal processing, instrumentation, and the development of new probing techniques. This image modeling technique can also be applied to other imaging modalities such as confocal and other laser scanning methods.

Mechanical and Optical Dynamic Model of Lung

A multiscale, multiphysics model generates synthetic images of alveolar compression under spherical indentation at the visceral pleura of an inflated lung. A mechanical model connects the millimeter scale of an indenter tip to the behavior of alveoli, walls, and membrane at the micrometer scale. A finite-difference model of optical coherence tomography (OCT) generates the resulting images. Results show good agreement with the experiments performed using a unique indenter-OCT system. The images depict the physical result with the addition of refractive artifacts and speckle. Compression of the alveoli alters the refractive effects, which introduce systematic errors in the computation of alveolar volume. The complete computational model is useful to evaluate new proposed imaging instrumentation and to develop algorithms for obtaining quantitative data on deformation. Among the potential applications, a better understanding of recruitment of alveoli during inflation of a lung, obtained through a combination of models and imaging could lead to improvements in noninvasive treatment of atelectasis.

Leveraging the crowd for annotation of retinal images

Medical data presents a number of challenges. It tends to be unstructured, noisy and protected. To train algorithms to understand medical images, doctors can label the condition associated with a particular image, but obtaining enough labels can be difficult. We propose an annotation approach which starts with a small pool of expertly annotated images and uses their expertise to rate the performance of crowd-sourced annotations. In this paper we demonstrate how to apply our approach for annotation of large-scale datasets of retinal images. We introduce a novel data validation procedure which is designed to cope with noisy ground-truth data and with non-consistent input from both experts and crowd-workers.

Automated retinal imaging and trend analysis – a tool for health monitoring

Most current diagnostic devices are expensive, require trained specialists to operate and gather static images with sparse data points. This leads to preventable diseases going undetected until late stage, resulting in greatly narrowed treatment options. This is especially true for retinal imaging. Future solutions are low cost, portable, self-administered by the patient, and capable of providing multiple data points, population analysis, and trending. This enables preventative interventions through mass accessibility, constant monitoring, and predictive modeling.

Mobile phone based mini-spectrometer for rapid screening of skin cancer

We demonstrate a highly sensitive mobile phone based spectrometer that has potential to detect cancerous skin lesions in a rapid, non-invasive manner. Earlier reports of low cost spectrometers utilize the camera of the mobile phone to image the field after moving through a diffraction grating. These approaches are inherently limited by the closed nature of mobile phone image sensors and built in optical elements. The system presented uses a novel integrated grating and sensor that is compact, accurate and calibrated. Resolutions of about 10 nm can be achieved. Additionally, UV and visible LED excitation sources are built into the device. Data collection and analysis is simplified using the wireless interfaces and logical control on the smart phone. Furthermore, by utilizing an external sensor, the mobile phone camera can be used in conjunction with spectral measurements. We are exploring ways to use this device to measure endogenous fluorescence of skin in order to distinguish cancerous from non-cancerous lesions with a mobile phone based dermatoscope.