Skip to main content

Alarm Verification

Built-in verification tools including video clip review, live view access, and event context to validate alarm authenticity. GCXONE uses AI-powered analytics to automatically classify alarms, reducing false positives by 80-95%.

✅

Verification

AI-Powered Analysis

Overview​

Alarm Verification is the process of determining whether an alarm represents a real security threat or a false positive. GCXONE provides multiple verification tools to help operators make accurate decisions quickly.

The Verification Process​

When an alarm occurs, GCXONE extracts three critical images (pre-alarm, current, and post-alarm) or a video clip. The AI engine classifies the object (Human or Vehicle) and places a bounding box around the threat.

AI-Powered Classification​

The core value of GCXONE is its AI-powered false alarm filtering, which reduces the volume of alerts by approximately 80% to 95%.

Classification Process:

  1. Image Extraction: System extracts key frames from video
  2. AI Analysis: AI engine scans for human or vehicle activity
  3. Bounding Box: If person detected, marks bounding box on image
  4. Classification: Determines if alarm is "Real" or "False"
  5. Automated Action: False alarms can be automatically closed

Verification Result Types​

  • Real Alarm: Confirmed security threat requiring response
  • False Alarm: Non-threatening event (e.g., shadows, animals, weather)
  • Technical Alarm: System or device issue requiring technical attention
  • Unverified: Unable to determine (requires manual review)

Quad View - The Power of Context​

For video events, the system displays three critical images: Pre-Alarm, Current-Alarm, and Post-Alarm. This "Quad View" provides complete context of what happened before, during, and after the alarm trigger.

Pre-Alarm Image​

Shows the scene before the alarm was triggered. This helps operators understand:

  • Normal state of the area
  • What changed to trigger the alarm
  • Environmental conditions

Current-Alarm Image​

Shows the scene at the moment the alarm was triggered. This is the primary verification image showing:

  • The detected object or activity
  • AI bounding box (if classification available)
  • Alarm trigger point

Post-Alarm Image​

Shows the scene after the alarm was triggered. This helps operators see:

  • If the threat persisted
  • Direction of movement
  • Current state of the area

AI Bounding Boxes​

If the AI has classified the object, you will see a bounding box around the person or vehicle:

  • Blue Box (Tracked): Object is being tracked but not necessarily a threat
  • Red Box (Alarm): Object has triggered an alarm condition
  • Green Box (Verified): Object has been verified as a real threat

Understanding Bounding Boxes​

  • Accuracy Indicator: Box confidence level
  • Object Type: Human, Vehicle, or Unknown
  • Movement Direction: Arrow showing direction of travel
  • Size Estimation: Relative size of detected object

Live View Access​

You can toggle to Live View to see what is happening now in real-time. This is useful for:

  • Active Threats: Monitor ongoing security incidents
  • Verification: Confirm current state of the area
  • Response Coordination: Guide security response in real-time

Live View Features​

  • Real-Time Streaming: Current camera feed
  • PTZ Control: Pan, tilt, zoom to investigate
  • Audio: Two-way audio communication (if available)
  • Recording: Start manual recording if needed

Archive Review​

Use the Timeline to scrub through the archive for further evidence. This allows you to:

  • Review History: See what happened before the alarm
  • Find Patterns: Identify recurring issues
  • Gather Evidence: Collect video evidence for reports
  • Investigate: Conduct thorough investigation of incidents

Timeline Navigation​

  • Scrub Bar: Drag to navigate through time
  • Thumbnail View: See key moments in timeline
  • Playback Controls: Play, pause, speed control
  • Export: Download video clips for evidence

Automated Filtering​

If the AI determines an alarm is "False" (e.g., caused by shadows or animals), it can be automatically closed without operator intervention. This significantly reduces operator workload.

Auto-Close Conditions​

  • High Confidence False: AI is very confident it's a false alarm
  • Pattern Recognition: Matches known false alarm patterns
  • Environmental Factors: Weather, lighting, or other environmental causes
  • Device Issues: Camera obstruction or technical problems

Manual Override​

Operators can always override AI classification:

  • Mark as Real: Even if AI says false
  • Mark as False: Even if AI says real
  • Request Review: Flag for supervisor review

Verification Workflow​

Step 1: Initial Assessment​

When an alarm appears in your queue:

  1. Review AI Classification: Check if AI has already classified the alarm
  2. View Quad Images: Review pre, current, and post-alarm images
  3. Check Bounding Boxes: See what the AI detected
  4. Read Alarm Details: Review site, device, and alarm type information

Step 2: Visual Verification​

  1. Open Live View: Check current state of the area
  2. Review Archive: Scrub through timeline if needed
  3. Compare Images: Compare pre-alarm vs. current state
  4. Look for Context: Check surrounding cameras if available

Step 3: Decision Making​

Based on verification:

  • Real Threat: Proceed with response workflow
  • False Alarm: Mark as false and close
  • Technical Issue: Escalate to technical team
  • Uncertain: Request supervisor review

Best Practices​

  1. Trust but Verify: Use AI classification as a guide, but always verify
  2. Use Context: Review all three quad images for complete picture
  3. Check Multiple Angles: If multiple cameras available, check all views
  4. Document Findings: Add notes explaining your verification decision
  5. Learn Patterns: Note common false alarm causes for future reference

Common False Alarm Causes​

  • Environmental: Shadows, reflections, weather
  • Animals: Wildlife triggering motion sensors
  • Vegetation: Trees, bushes moving in wind
  • Lighting: Sudden light changes
  • Technical: Camera obstructions, lens issues

Need Help?​

If you're experiencing issues with alarm verification, check our Troubleshooting Guide or contact support.