Colorimetric assays are an important tool in point-of-care testing that offers several advantages to traditional testing methods such as rapid response times and inexpensive costs. A factor that currently limits the portability and accessibility of these assays are methods that can objectively determine the results of these assays. Current solutions consist of creating a test reader that standardizes the conditions the strip is under before being measured in some way. However, this increases the cost and decreases the portability of these assays. The focus of this study is to create a machine learning algorithm that can objectively determine results of colorimetric assays under varying conditions. To ensure the flexibility of a model to several types of colorimetric assays, three models were trained on the same convolutional neural network with different datasets. The images these models are trained on consist of positive and negative images of ETG, fentanyl, and HPV Antibodies test strips taken under different lighting and background conditions. A fourth model is trained on an image set composed of all three strip types. The results from these models show it is able to predict positive and negative results to a high level of accuracy.
First, a pneumotach based flow sensing technique has been developed and integrated into a face mask for respiratory profile tracking. Algorithms have been developed to convert the pressure profile into respiratory flow rate profile. Gyroscope-based correction is used to remove motion artifacts that arise from daily activities. By using Principal Component Analysis, the follow-up work established a unique respiratory signature for each subject based on the flow profile and lung parameters computed using the wearable mask system.
Next, wristwatch devices to track transcutaneous gases like oxygen (TcO2) and carbon dioxide (TcCO2), and oximetry (SpO2) have been developed. Two chemical sensing approaches have been explored. In the first approach, miniaturized low-cost commercial sensors have been integrated into the wristwatch for transcutaneous gas sensing. In the second approach, CMOS camera-based colorimetric sensors are integrated into the wristwatch, where a part of camera frame is used for photoplethysmography while the remaining part tracks the optical signal from colorimetric sensors.
Finally, the wireless connectivity using Bluetooth Low Energy (BLE) in wearable systems has been explored and a data transmission protocol between wearables and host for reliable transfer has been developed. To improve the transmission reliability, the host is designed to use queue-based re-request routine to notify the wearable device of the missing packets that should be re-transmitted. This approach avoids the issue of host dependent packet losses and ensures that all the necessary information is received.
The works in this thesis have provided technical solutions to address challenges in wearable technologies, ranging from chemical sensing, flow sensing, data analysis, to wireless data transmission. These works have demonstrated transformation of traditional bench-top medical equipment into non-invasive, unobtrusive, ergonomic & stand-alone healthcare devices.
For each of these cases, I present a feasible image restoration pipeline to correct for their particular limitations. For the pinhole camera, I present an early pipeline to allow for practical pinhole photography by reducing noise levels caused by low-light imaging, enhancing exposure levels, and sharpening the blur caused by the pinhole. For lensless cameras, we explore a neural network architecture that performs joint image reconstruction and point spread function (PSF) estimation to robustly recover images captured with multiple PSFs from different cameras. Using adversarial learning, this approach achieves improved reconstruction results that do not require explicit knowledge of the PSF at test-time and shows an added improvement in the reconstruction model’s ability to generalize to variations in the camera’s PSF. This allows lensless cameras to be utilized in a wider range of applications that require multiple cameras without the need to explicitly train a separate model for each new camera. For UDCs, we utilize a multi-stage approach to correct for low light transmission, blur, and haze. This pipeline uses a PyNET deep neural network architecture to perform a majority of the restoration, while additionally using a traditional optimization approach which is then fused in a learned manner in the second stage to improve high-frequency features. I show results from this novel fusion approach that is on-par with the state of the art.
This research project focuses on analyzing the variance in shielding effectiveness and electromagnetic field effects of a thin film of Cobalt Zirconium Tantalum Boron (CZTB) in the band of frequencies most likely to require innovative solutions to long-standing problems of noise and interference. The measurements include Near H-Field attenuation and field effects, Far Field shielding, and Backscatter. Minor variances in the thickness and layering of sputter deposition can have significant changes electromagnetic signature of devices which radiate energy through the material.
The material properties presented in this research are H-Field attenuation, H-Field Flux Orientation, Far-Field Approximation, E Field Vector Directivity, H Field Vector Directivity, and Backscatter Magnitude. The results are presented, analyzed and explained using characterization techniques. Future work includes the effect of sputter deposition orientation, application to devices, and applicability in mitigating specific noise signals beyond the 5G band.