[Technical Notice] GV-AI Guard PVD Motion Detection Usage and Case Studies

[joycechang]

Article ID: GV2-23-06-2
Release Date: June 2, 2023

Applied to
GV-AI Guard V1.1

Introduction
False alarms have long plagued security personnel, so reducing false alarms is a top priority for surveillance systems. People and Vehicles Motion Detection (PVD) in GV-AI Guard uses deep-learning algorithms to provide more robust and accurate motion detection capabilities than other motion detection approaches.

Although PVD motion detection has significantly advanced motion detection capabilities, achieving 100% accuracy is highly challenging due to inherent complexities and uncertainties in real-world scenarios. The PVD algorithms use several detection thresholds to determine detection sensitivity or to filter out false positives. You can fine-tune the thresholds to optimize the accuracy and performance of PVD motion detection for your cameras and applications.

The documentation also includes case studies on how to fine-tune the PVD motion detection, as well as the most recent software patch for improving the current PVD motion detection.

Pls. find the documentation at this link: https://s3.amazonaws.com/geovision_down ... tudies.pdf

Latest Software Patch
With the latest patch V1.1.0.2, you can now adjust the detection thresholds. The patch is only applicable to GV-AI Guard V1.1.

GV-AI Guard patch V1.1.0.2 download: https://dlcdn.geovision.com.tw/Software ... .1.0.2.zip