Methods for 2D and 3D Endobronchial Ultrasound Image Segmentation
We propose computer-based methods for segmenting 2D EBUS (Endobronchial ultrasound) frames and 3D EBUS sequences. Both segmentation methods compare very favorably to ground-truth results for a human lung-cancer patient EBUS database. We also demonstrate the potential of the methods for EBUS localization in a multimodal image-guided bronchoscopy system.
QRS Detection Algorithm for Telehealth Electrocardiogram Recordings
We describe a QRS detection algorithm that is suitable for clean clinical ECGs as well as poorer quality telehealth ECG. A data repository of annotated telehealth ECGs has been made available on-line for future algorithmic development and testing.
Towards a Portable Cancer Diagnostic Tool Using a Disposable MEMS-based Biochip
A portable cancer diagnostic tool integrated with a disposable MEMS-based biochip is developed for measuring electro-thermo-mechanical (ETM) properties of the breast tissue. The ETM properties of the normal and cancerous breast tissues are measured by indenting the tissue placed on the biochip integrated inside the 3D printed tool.
Biomimetic Accommodating Intraocular Lens Using a Valved Deformable Liquid Balloon
Presbyopia is a common age-related condition that prevents people from focusing on near objects due to a hardened lens. To address this problem we designed a unique lens that was a flexible, inflatable balloon with a valve. This lens allowed the eye to use its natural focusing muscles to restore near and far vision.
Methods for Improving the Curvature of Steerable Needles in Biological Tissue
Robotic systems that can insert flexible needles along curved paths through the body have the potential to improve a variety of percutaneous interventions, including ablation of liver tumors. This paper describes finite-element modeling and experiments in liver tissue samples, which demonstrate that optimization of steerable needle tip geometry can greatly improve needle curvature in liver tissue. A new articulated-tip steerable needle is also described, which allows steerable needles with exaggerated tip geometry to be deployed in realistic pre-clinical testing.
Epileptic Seizure Prediction Based on Multivariate Statistical Process Control of Heart Rate Variability Features
We proposed a new epileptic seizure prediction method utilizing heart rate variability (HRV) analysis. It monitors time-frequency-domain HRV features for predicting seizures by using multivariate statistical process control (MSPC). The application results to clinical data produced accurate predictions (91%) for epileptic seizures and there were few false-positives (0.7 times/hour). The possibility of realizing a HRV-based epileptic seizure prediction system was shown.