| I&T Solution |
Detecting Intruders & Abnormal Behaviors
(REF: S-1403) |
| Trial Project |
|
| Solution Feature |
- Comply ALL requirements and DO More e.g our camera-based Visual AI system can differentiate vehicles from human & workers from intruders. It can detect intruders’ abnormal behaviors 24x7 alerting for any act of vandalism, arson, climbing up, abandoning objects, graffiti, sleeping over, removing/opening up of tunnel settings & more).
- Real-time Alert can be sent as an email to a designated mobile device or a workstation in the control room informing 1st Responder/Supervisor. The Alert is associated with a 20-second of jump-to-event video clip (10-second of pre and post alarm), helping supervisor to understand the entire situation facilitating remedial action.
- We can differentiate vehicles from human by its shape and characteristics, differentiate workers from intruders by their PPE/uniforms or making use of their secondary id (such as mobile phone or long-distance badge card or QR code etc). We have 300+ ready made AI use cases that can ensure quick deployment.
- We can record any human’s (either workers or intruders) in and out attendance record by placing the cameras at tunnel entrance and exit. Workers in-and-out-time and their attendance are recorded respectively upon entry and upon exit. (even by using low resolution or any standard CCTV camera)
- An alarm will be triggered as soon as an intruder is detected entering the tunnel. Subsequent alarm will be sent again based on his behaviors that could be a safety or hazardous concern such as graffiti, sabotaging the equipment room or road signs and reopening tunnel tube etc.
|
| Trial Application and Expected Outcome |
- Trial can be conducted immediately without the need to supply ANY datasets for deep learning purpose or your expensive investment in computational equipment. (Refer to our unique Hybrid AI approach) Typical trial that will make use of our existing AI modules can be done within a 1 week.
- We will produce a trial plan that will allow you to measure the performance of our Analytics through three crucial factors: Detection Accuracy, False Alarm Rate and Missed Detection. Our AI is equipped with Nuisance Alarm Minimization System (NAMS) that can differentiate real intruders from other false alarms.
- Our Analytics system can be setup to run on a notebook or edge computer & it can easily ingest and analyze either using Live or recorded video. To make the trial easier and more precise, we will stage the most realistic scenarios & record them with you.
- We will prepare you a test plan of how to stage each possible scenarios seeking for your approval before. These scenarios will include the followings: an intruder entering from one view and exiting from the other view; a person sleeping over, abandoning an object, sabotage or remove a signboard, etc.
- Additionally you can supply us with more scenarios to test out. You can record our AI performance based on the given assessment sheet/score sheet.
|
| Additional Solution Information |
WildFaces Multi-Sensory AI for EMSD - intrusion.pdf
|
| Info on I&T Solution Provider |
| Solution Provider | : | WildFaces Technology Limited | | Address | : | Unit 22, Level 6, Core C, Cyberport 3, 100 Cyberport Road, Hong Kong | | Contact Person | : | Eric Wong |
| Position | : | Senior AI Business Consultant | | Tel | : | +852 8216 2219 | | Email | : |
bd@wildfaces.ai | | Webpage | : | www.wildfaces.ai |
|
