Run Detection from Existing CCTV

I&T Solution Run Detection from Existing CCTV
(REF : S-1312)
Trial Project
Solution Feature
  • Detect 17 human body key points, including nose, eyes, ears, shoulders, elbows, wrist, hip, knees etc. to identify running action
  • Provide custom abnormal behavior alerts with sound on the UI dashboard and SMS to the duty station controller when the AI model reaches a certain threshold
  • Integrate with the Public Address System (PAS) to play an announcement to the passenger at the concerned area
  • Provide incident statistics dashboard for users to interactively view incident statistics recorded in the system and generate statistics report with graphs and charts
  • Allow users to report false alert (for future model retraining), report missed incident and add a comment to a incident record
Trial Application and Expected Outcome
  • Increase efficiency in detecting running and lower accident rate by 80% caused by running passengers by taking shorter than 0.5s to detect skeleton for image frame, and send SMS alert
  • Integrate with the Public Address System (PAS) to play an announcement to the passenger at the concerned area by users’ confirmation on the relevant image or video
  • Keep record of detected incident list with details including time, location and replay of the detected running passengers for further analysis
  • Visualize incident statistics by generating graphs and charts automatically, and allow users to generate relevant statistics report for reporting to management
  • Achieve more than 80% of accuracy rate in detecting running passengers and target to improve the accuracy by retraining data periodically
Info on I&T Solution Provider
Solution Provider:Datax Limited
Address:Unit 707-708, 7/F, Data Technology Hub (DT Hub), 5 Chun Cheong Street, Tseung Kwan O
Contact Person:Aric Fung
Position:Account Director
Tel: 91438473
Email: aric.fung@datax.io
Webpage: https://datax.io/

For details of the above I&T solution, please contact the I&T solution provider.