Technology Overview

Absence Detection is an AI-based video analytics technology designed to determine whether personnel are present at designated posts, workstations, service counters, or other locations where continuous or scheduled attendance is expected.

Typical monitored locations can include reception desks, security guard posts, entrance gates, service counters, production stations, control desks, and other operational positions.

Unlike many video analytics technologies that detect the presence of a particular object or behavior, Absence Detection identifies the absence of an expected person.

The technology combines AI-based person detection with a predefined monitoring zone and configurable absence duration. When no person is detected within the designated area for longer than the permitted period, the system generates an Absence Detection event.

How Absence Detection Works

The AI continuously processes video from the surveillance camera and detects people within the monitored scene.

A specific area representing the required work post is configured as an Absence Detection zone. The analytics system then continuously evaluates whether a person is present within this zone.

The basic detection logic can be represented as:

Person present → normal status

Person leaves → absence timer begins

Person returns before threshold → timer resets

Person remains absent beyond threshold → Absence Detection event

This combination of spatial and temporal analysis prevents the system from generating unnecessary alarms every time an employee briefly moves away from the exact working position.

AI-Based Person Detection

Person detection provides the foundation of Absence Detection.

The AI analyzes the video image and identifies visual characteristics associated with the human body. Detected people can then be evaluated relative to the configured monitoring zone.

For Absence Detection, the system does not necessarily need to determine the person's identity. Its primary objective is to determine whether the required area is occupied by a person or unattended.

This makes the technology suitable for applications where the presence of personnel is important but individual identification is unnecessary.

Monitoring Zones

Correct configuration of the monitoring zone is one of the most important elements of Absence Detection.

The zone should represent the area where personnel are reasonably expected to be visible while performing their duties.

For example, at a reception desk, the monitoring zone might include the chair and normal working area behind the counter rather than only one precise position.

At a security gate, the zone might cover the guard booth or designated observation area.

At a production station, the zone can correspond to the area where an operator normally stands or sits while operating the equipment.

Providing sufficient space for normal movement helps prevent an employee from being considered absent simply because they shifted position within their workstation.

Configurable Absence Duration

The time threshold is a key component of the technology.

In most environments, personnel may legitimately leave their exact position for short periods. A receptionist may retrieve a document, a security guard may briefly inspect something nearby, or an operator may move within the surrounding work area.

Generating an event immediately after a person disappears from the detection zone would therefore create unnecessary alarms.

Instead, Absence Detection uses a configurable duration threshold.

For example, if the absence threshold is configured to 60 seconds, the system begins measuring time when the monitored zone becomes unoccupied. If a person returns within 60 seconds, no absence event is generated.

If the zone remains unattended beyond the configured threshold, the system generates an event.

This allows organizations to define what constitutes a meaningful absence according to the requirements of each monitored post.

Presence and Absence State Analysis

Absence Detection can be understood as a state-based analytics process.

The monitored area transitions between two principal states:

OCCUPIED — a person is detected within the configured zone.

UNOCCUPIED — no person is currently detected within the configured zone.

Time is then applied to the unoccupied state to determine whether it should become an actionable absence event.

This approach provides greater reliability than treating every disappearance of a person as an alarm condition.

Returning to the Work Post

When a person returns to the monitored zone, the analytics system detects the renewed presence and resets or terminates the absence condition.

Depending on system configuration, the return of personnel may also be recorded as the end of the event.

This makes it possible for the surveillance system to establish an event timeline such as:

14:32:10 — Work post becomes unattended

14:33:10 — Configured absence threshold exceeded

14:33:10 — Absence event generated

14:36:25 — Person returns to monitored post

Such information can assist operators with both real-time monitoring and later investigation.

Identity-Independent Detection

Absence Detection does not inherently determine whether the person occupying the workstation is the specifically assigned employee.

For example, if a receptionist leaves and another person remains within the configured reception zone, basic Absence Detection may continue to consider the post occupied.

The technology therefore answers the question:

“Is a person present at this post?”

rather than:

“Is the specifically assigned person present at this post?”

Applications requiring verification of a particular employee would require additional identification technologies or integration with other systems.

This distinction is important when designing and configuring the application.

Event Generation and Integration

Once the monitored zone remains unattended beyond the configured duration, the analytics system generates an Absence Detection event.

The event can be integrated into the broader ACTi video surveillance workflow and may be used to:

Display an alert to an operator or supervisor Highlight the corresponding surveillance camera Trigger or bookmark video recording Generate an alarm or notification Create an event record for later search Provide associated video for verification Initiate other configured operational workflows

This enables supervisors to focus on actual exceptions rather than continuously observing multiple work posts through live video.

Camera Placement and Image Requirements

Reliable person detection depends significantly on camera placement and image quality.

The camera should provide a clear and consistent view of the monitored work post and the area in which personnel normally perform their duties.

Important installation considerations include:

Clear visibility of the monitored workstation Sufficient person size within the image Appropriate camera viewing angle Adequate and consistent illumination Suitable video resolution and compression quality Minimal obstruction by furniture, equipment, doors, or other objects Proper coverage of normal personnel movement within the workstation

The camera should be positioned so that a person performing normal duties remains sufficiently visible to the AI.

Occlusion Considerations

Occlusion is particularly important for Absence Detection because an obscured person could potentially appear to the analytics system as an empty workstation.

For example, a receptionist may be partially hidden behind a high counter, a security guard may be obscured by the structure of a booth, or a production operator may move behind machinery.

If the AI cannot obtain sufficient visual evidence of the person, the monitored area may temporarily appear unoccupied.

Camera positioning and detection-zone design should therefore minimize predictable obstructions.

Seated Personnel

Some applications involve employees who spend much of their working time seated.

A seated person may present different visual characteristics from a standing or walking person, particularly when a desk or counter obscures the lower body.

The camera should therefore provide enough visibility of the person's upper body for reliable detection.

Reception desks and service counters should be evaluated carefully because high counters, computer monitors, partitions, and furniture can significantly reduce person visibility.

Multiple People

The primary requirement for basic Absence Detection is typically the presence of at least one person within the designated zone.

If several people enter the monitored area, the post remains occupied as long as the required presence condition is satisfied.

The disappearance of one individual does not necessarily generate an absence event if another person remains within the monitored zone.

Applications requiring a specific number of personnel or verification of specific individuals represent a different analytical requirement and may require additional technologies.

Environmental Factors

Several environmental conditions can influence detection performance, including:

Poor or changing illumination Strong backlighting Shadows Low-resolution video Excessive video compression Partial body occlusion Reflections or transparent partitions Very small subjects within the image Unusual camera perspectives Changes to furniture or workstation layout

Detection performance should therefore be evaluated under representative operating conditions.

Avoiding False Absence Events

Proper system configuration can significantly reduce unnecessary absence alerts.

Important considerations include:

Defining a sufficiently large monitoring zone for normal working movement Selecting an appropriate absence duration Providing clear camera visibility of seated and standing personnel Avoiding areas with frequent visual obstruction Accounting for legitimate short-term movement away from the post Testing the configuration during normal daily operations

An excessively small detection zone or an excessively short absence threshold can produce unnecessary events even when employees are performing their duties normally.

Real-Time Processing

Absence Detection continuously evaluates the occupancy state of the configured monitoring area.

Person detection itself can occur continuously, while the configured absence timer determines when the unoccupied condition becomes an actionable event.

The effective alert time can therefore be represented conceptually as:

Person leaves monitored zone + configured absence duration → Absence Detection alert

This allows system administrators to prioritize either rapid notification or greater tolerance for brief legitimate absences according to the application.

Privacy and Identity

Absence Detection can operate without facial recognition or other biometric identification.

The analytical objective is to determine whether personnel are present within a designated work area, rather than determine who the individual is.

This identity-independent approach allows the technology to provide automated operational monitoring while limiting the amount of personal information required by the analytics process.

Organizations should nevertheless deploy the technology in accordance with applicable workplace monitoring, privacy, labor, and data-protection requirements.

Technical Benefits

AI-powered Absence Detection transforms surveillance cameras into automated workstation-presence monitoring devices.

Key technical advantages include:

AI-based person detection Continuous workstation occupancy monitoring Configurable monitoring zones Configurable absence duration Automatic filtering of brief absences Real-time event generation Integration with recording, alarms, and notifications Event-based video search and investigation Monitoring of multiple posts without continuous operator observation Operation without requiring personal identification Performance Considerations

Absence Detection is an AI-assisted technology, and its reliability depends on the system's ability to consistently detect personnel within the configured area.

Poor visibility, occlusion, unsuitable camera angles, insufficient subject size, lighting changes, or inappropriate detection-zone configuration can cause a person to temporarily become undetectable and potentially create a false absence condition.

Conversely, another person entering the monitoring zone may cause a basic presence-based system to consider the workstation occupied even if the originally assigned employee is absent.

System designers should therefore carefully consider what “presence” means for the intended application and configure the camera view, monitoring zone, and absence threshold accordingly.

Testing and tuning under representative working conditions is recommended before deployment.

Conclusion

AI-powered Absence Detection enables surveillance systems to automatically determine when designated work posts become unattended and convert prolonged absence into actionable operational events.

By combining AI-based person detection, configurable monitoring zones, occupancy-state analysis, and time-based thresholds, the technology can distinguish meaningful absences from ordinary short-term movement around a workstation.

When deployed with suitable camera positioning, appropriate zone configuration, and carefully selected absence-duration thresholds, Absence Detection provides an effective way to monitor reception desks, security posts, service counters, production stations, and other locations where personnel presence is important.

The automatic absence detection is very helpful for both business and security management.

The security guard leaving the post, or missing officers at the security checkpoints in the airports, would be a serious security risk. The automatic alert can help the supervisors handle the situation immediately.

Missing employees in the factory production line would interrupt the production flow. Missing employees at the check-out counters or at the reception desks would be bad for the business as the customers would have to wait and in some cases even leave without the purchase.

It is possible to use ACTi's Absence Detection cameras together with ACTi's ZNR-series NVR or with Windows-based NVR 3 servers. The VMS will receive the absence detection alerts from the cameras and handle the events by video pop-up, sound alert, sending an email alert and more.

Furthermore, some of the latest ZNR-series NVR servers have the built-in Absence Detection capability - it is possible to use a regular surveillance camera without built-in AI and still run the absence detection thanks to the NVR-side AI capabilities.
Read More