Steven D. Schwartz

Ali Hatamizadeh, Hamid Hosseini, Niraj Patel et al.

The retinal vasculature provides important clues in the diagnosis and monitoring of systemic diseases including hypertension and diabetes. The microvascular system is of primary involvement in such conditions, and the retina is the only anatomical site where the microvasculature can be directly observed. The objective assessment of retinal vessels has long been considered a surrogate biomarker for systemic vascular diseases, and with recent advancements in retinal imaging and computer vision technologies, this topic has become the subject of renewed attention. In this paper, we present a novel dataset, dubbed RAVIR, for the semantic segmentation of Retinal Arteries and Veins in Infrared Reflectance (IR) imaging. It enables the creation of deep learning-based models that distinguish extracted vessel type without extensive post-processing. We propose a novel deep learning-based methodology, denoted as SegRAVIR, for the semantic segmentation of retinal arteries and veins and the quantitative measurement of the widths of segmented vessels. Our extensive experiments validate the effectiveness of SegRAVIR and demonstrate its superior performance in comparison to state-of-the-art models. Additionally, we propose a knowledge distillation framework for the domain adaptation of RAVIR pretrained networks on color images. We demonstrate that our pretraining procedure yields new state-of-the-art benchmarks on the DRIVE, STARE, and CHASE_DB1 datasets. Dataset link: https://ravirdataset.github.io/data/

Fereshteh Shahmiri, Steven Schwartz, Can Usanmaz

This work presents a triboelectric nanogenerator-based (TENG) therapeutic stress ball to provide gesture monitoring and physiological data on patients requiring physical therapy of various degrees. The device utilizes a 5-layer stack of silicone and braided silver-coated nylon rope electrodes to create a sensor network that monitors 40-points across the surface of a semi-spherical prototype. A modified version of a standard ECG circuit was utilized to provide proper loading, noise rejection, filtering, and phase of the TENG signals along with multiplexing of the many electrodes. All system components were selected with a final embedded system in mind. Testing of the device was conducted utilizing an Arduino Uno and an EVAL-AD5940BIOZ evaluation board for electrodermal activity for stress and/or pain after exercise. An accelerometer was included for device activation and hand tremor detection. Upon testing, the self-powered TENG sensors produce positive impulses upon contact and negative impulses upon release of contact from the surface of the ball. Furthermore, finger removal detection was demonstrated by capturing the associated negative impulse by maintaining the bipolar signal in our conditioning circuit. EDA results indicate silver-coated nylon as a potentially good dry-electrode which can be used with even more electrodes for bio-impedance or ECG capture to further expand the device functionality. A MATLAB-based GUI was designed to provide the user with data tracking and visual monitoring of the data via serial communication from the microcontrollers. Finally, it should be noted that this provides a means for low-cost low-power gesture tracking without the use of flexible capacitive grid arrays and provides the user with a pleasant tactile experience that one expects form a stress ball due to its unique material design.