Remark AI Safety Space Platform (SSP) enables existing CCTV with powerful AI features, detecting and recognizing faces, pedestrian behavior, vehicle traffic, abnormal object, smoke, and fire.

Jump to

Remark AI Safety Space Platform Solution Overview

Remark AI Safety Space Platform (SSP) enables existing CCTV with powerful AI features, detecting and recognizing faces, pedestrian behavior, vehicle traffic, abnormal object, smoke, and fire.

Key market application

Metasearch across cameras all over the site and city for police and security to trace targets (e.g., fugitive, suspicious, lost child)

Analytic models are applied for all the detected faces, pedestrians, and vehicles to extract characteristics, e.g., appearance similarity, age, gender, wearing style, color, and plate number.

Even with fragmented information by the witnesses (e.g., age range, partial plate number), the system can find the targets at once. Additional time and location filters enhance the accuracy. Searched results can help to reconstruct the temporal and spatial historical track and most recent location to organize the operation.

Real-time alert for the operator to monitor a large number of cameras 24/7 with a focus on an important event

Video analytics detect and track faces, pedestrians, vehicles, and abnormal objects and perform high-level behavior patterns to recognize events, e.g., intrusion, loitering, falling, illegal parking, driving in a reverse direction, unattended objects, etc. Real-time alerts with notifications are sent to the system and mobile App for operators to verify with large images, video clips, or live video so that operators can handle only suspicious events and take action at once, to avoid losing attention for long silent periods.

High-level event recognition, including unattended bags, etc.

Unattended objects cause a wide range of concerns for public administrators, e.g., explosives from terrorists, mischief, belonging of careless passengers, theft, and fly-tipping.

Model alerts abnormal objects to operators to evaluate the object's risk and react quickly. The system also traces back the pedestrian and vehicle recently connected to it previously to find the abandoning handler.

Combined with our metasearch function above, a targeting handler can easily be traced from the current location and the historical track.

Features and benefits

Safety Space Platform is an IoT platform performing real-time face, pedestrian, vehicle, and object analytics on existing security cameras with low deployment requirements.

It covers a wide range of AI video analytics features:

Face : 

Facial ID matching and alert on watchlist; and similarity search, face attribute search: demographic attribute, mask wearing;

Pedestrian : 

pedestrian intruders, loiterers, leavers, crowds, in/out flow, attribute search: demographic attire style and color, PPE, and falling down behavior. 

Vehicle : 

vehicle intruder, parking, leaving, reverse direction, traffic flow, ANPR and search. 

Abnormal unattended object :

Last contacted pedestrian/vehicle tracking. 

 

Safety Space Platform is also a one-stop platform to manage devices and event history. Live streaming surveillance from multiple sites. 

Event alert with map view visualization and export, alert flexible time slot, and email/SMS notification with fast reaction.

History Search for various combinations of event attribute criteria to trace the target at once (similarity, wearing, number plate, etc.). 

Dashboard to review overview analytics for areas and alert. 

Solution Architecture Diagram

The Safety Space Platform processes video stream and extracts the information from video stream. The extracted information is an event type, timestamp, etc. The event type can be Pedestrian Intrusion, Face Recognition, Vehicle Intrusion with ANPR, Unattended Object detection, and PPE detection. The Safety Space Platform includes additional information about VMS name, VMS location. The extracted information and additional information will be transferred to the GSC by using Genetec SDK/API. The transferred information will be natively recognized by GSC as a Bookmark.

Integration with GSC Architecture : 

Communication between GSC and Safety Space Platform is established by using GSC REST API. A service is taking care of periodically download the camera list from GSC and push records back to GSC.

This service is run by an additional hardware to be installed in the same network as GSC. This Edge AI Box is provided by REMARK AI and it's also in charge of running the video analytics.