AR-Assisted Sacral Neuromodulation
AR-Assisted Sacral Neuromodulation
AR-Assisted Sacral Neuromodulation
AR-Assisted Sacral Neuromodulation
Organization:
UC San Diego
Organization:
UC San Diego
Organization:
UC San Diego
Duration:
Mar 2023 – Now
Duration:
Mar 2023 – Now
Duration:
Mar 2023 – Now
Video Demo
Overview
Designed and 3D-printed target models for tracking; implemented a benchmark method using MRTK and Vuforia SDK.
Streamlined computer vision-based detection and tracking by training custom YOLOv4 weights with Anaconda and Tensorflow, integrating Deep SORT for real-time tracking, realizing seamless data transfer between Python and Unity.
Enhanced tracking capabilities by leveraging visible-light cameras, infrared cameras, and depth cameras from HoloLens2 Research Mode. Achieved a 31.9% improvement in precision and a 52.6% reduction in latency.
Video Demo
Overview
Designed and 3D-printed target models for tracking; implemented a benchmark method using MRTK and Vuforia SDK.
Streamlined computer vision-based detection and tracking by training custom YOLOv4 weights with Anaconda and Tensorflow, integrating Deep SORT for real-time tracking, realizing seamless data transfer between Python and Unity.
Enhanced tracking capabilities by leveraging visible-light cameras, infrared cameras, and depth cameras from HoloLens2 Research Mode. Achieved a 31.9% improvement in precision and a 52.6% reduction in latency.
Video Demo
Overview
Designed and 3D-printed target models for tracking; implemented a benchmark method using MRTK and Vuforia SDK.
Streamlined computer vision-based detection and tracking by training custom YOLOv4 weights with Anaconda and Tensorflow, integrating Deep SORT for real-time tracking, realizing seamless data transfer between Python and Unity.
Enhanced tracking capabilities by leveraging visible-light cameras, infrared cameras, and depth cameras from HoloLens2 Research Mode. Achieved a 31.9% improvement in precision and a 52.6% reduction in latency.
Video Demo
Overview
Designed and 3D-printed target models for tracking; implemented a benchmark method using MRTK and Vuforia SDK.
Streamlined computer vision-based detection and tracking by training custom YOLOv4 weights with Anaconda and Tensorflow, integrating Deep SORT for real-time tracking, realizing seamless data transfer between Python and Unity.
Enhanced tracking capabilities by leveraging visible-light cameras, infrared cameras, and depth cameras from HoloLens2 Research Mode. Achieved a 31.9% improvement in precision and a 52.6% reduction in latency.
