Title

Robust human target detection and acquisition

Description

Robust human target detection and acquisition/locking without losing track under occlusions for outdoor operational scenarios. Visuals alarm generation upon anomaly detection (i.e. human jumping/ crawling/ running)

Objectives

• Long-range Robust Human Detection
• The model should not lose focus even if the target is behind some obstacles
• Anomaly detection for the detected human
• Visual alarm if an anomaly is detected
• Multiple camera support for the service
• Offline working of the whole system
• Infrared and Night vision support
• Take multiple pictures of the intruder if any anomaly is detected

Resources

• ActionXPose model from the Newcastle University
  • Angelini F 2020.pdf (ncl.ac.uk)
  • [Pose‐driven human activity anomaly detection in a CCTV‐like environment - Yang - 2023 - IET Image Processing - Wiley Online Library](https://ietresearch.onlinelibrary.wiley.com/doi/pdf/10.1049/ipr2.12664#:~:text=Therefore%2C a human pose-drivenand object-detector-based deep learning architecture,RGB based counterpartlimitations in different human activities classification.)
  • 1810.12126v1.pdf (arxiv.org)
  • compilation - Is it possible to compile a program written in Python? - Stack Overflow
• Dashboard
  • How to Easily Build Desktop Apps with HTML, CSS and Javascript (freecodecamp.org)

Implementation Details