event-template-mapping

To map LLM visual observations to video surveillance event candidates, run the event-template-mapping script with JSON inputs containing frame observations, visual timelines, and camera IDs to generate event candidates with confidence and severity scores.

Event candidate semantic mapping in security monitoring converts raw LLM visual observations or method-level event results into structured event types, assigning confidence, severity, reason, and evidence frames for surveillance systems.

To convert method-level event results into structured event templates for camera feeds, execute the mapping script by passing the method results JSON, template name, camera ID, and configuration file to output categorized event semantics.

No, you cannot use this event mapping skill without a predefined event template name. The script execution requires a template name, camera ID, and configuration JSON to properly categorize visual observations into event candidates.

The JSON input format required for processing visual observations into event candidates must include frame observations, visual timelines, event templates, and camera IDs to accurately generate event candidates with confidence and severity metrics.

Video monitoring event mapping fails to generate severity and confidence scores when the input JSON lacks required frame observations or visual timeline data. The script needs complete method results and a valid template name to map event candidates successfully.