Intelligent Video Search Technology

IvedaAI™ uses deep learning technology for advanced AI Video Search, similar to Google for text search. AI Video Search allows users to search for a combination of objects (100+ types) with multiple colors and quantities in both real-time and recorded videos.

Our IvedaAI appliance works with any ONVIF-compliant IP cameras and most popular VMS platforms, enabling accurate search across dozens to thousands of cameras in less than 1 second. IvedaAI products are designed to maximize efficiency, save time, and cut cost.

Our appliance includes multiple AI functions based on customer needs.

Intelligent Video Search Technology

Build your own. Choose different AI functions in one appliance.

Intrusion Detection

  • Detect based on object types, colors, quantities, and time frame
  • Detect different types of objects in the same ROI
  • Draw multiple ROIs (region of interest) of any shape for different types of objects
  • Set different sensitivity levels to balance between detection speed and detection accuracy
  • Achieve high performance in daytime, nighttime, indoors and outdoors
  • Eliminate false alarms caused by irrelevant objects
  • Common use cases: perimeter protection, parking detection, vehicle zoning

Face Search

  • Detect face automatically from IP cameras or user-uploaded videos/images
  • Detect multiple faces simultaneously in one video/images
  • Enable quick search for missing person across multiple cameras
  • Locate people within seconds from uploaded photo/video
  • Can detect and match multiple faces in one image
  • Comply with privacy laws by not saving/requiring any ID information
  • Common use cases: locating missing person in shopping center, locating suspects on school campus

Face Recognition

  • Compare an uploaded face with a pre-configured face database and display similarity levels
  • Detect Approximate age range and gender
  • Enable quick search for missing person across multiple cameras
  • Locate people within seconds from uploaded photo/video
  • Can detect and match multiple faces in one image
  • Comply with privacy laws by not saving/requiring any ID information
  • Common use cases: locating missing person in shopping center, locating suspects on school campus

License Plate Recognition

  • Capture and read license plates at a slight angle (~20 degrees, not completely facing cameras)
  • Match detected license plates automatically with saved lists
  • Search for partial license plate number (i.e., enter any characters in license plates to search for potential matches in the databasep
  • Support three modes for different applications with different resource usage:
    • Parking Lot
    • City Road
    • Highway
  • Vehicle search based on type and color
  • Upload multiple blacklists / whitelists

Weapon Detection

  • Detect handguns, shotguns, rifles, and machine guns that appear in the camera’s field of view​
  • Provide critical video images to handle active shooter events in real time
  • Identify weapons in real time across multiple cameras (< 1 seconds)
  • Integrate with 3rd-party systems to trigger lockdowns upon detection

Smoke and Fire Detection

  • Detect smoke and fire in real time​
  • Receive alerts of detection instantaneously
  • Integrate with fire alarm systems to trigger alarm upon detection
  • Use in any size facility to maximize safety of students, employees and/or customers​
  • Use in conjunction with other AI analytics for complete safety

People Counting/Vehicle Counting

  • Count people and/or vehicle (multiple types) moving in various directions in real time
  • Enable user to define inbound vs. outbound direction for every counting line
  • Enable real-time alerts for abnormal events:
    • people/vehicle moving in the wrong direction
    • person falling/crouching
    • loitering
    • illegal parking
  • Support multiple counting lines per camera field of view
  • Provide statistics organized by camera, counting line, object type, time period
  • Provide business intelligence: peak hours, traffic flow, occupancy per unit time

Health Analytics

  • Temperature detection and alert
  • Social distancing detection
  • PPE Detection (e.g., mask, hard hat
  • Occupancy/Over-capacity detection

Heatmap

  • Generate heatmaps based on object detection
  • Show the frequency of appearance of the selected object types (e.g. car, person, animal)
  • Display statistics of gender and age groups
  • Display heatmap results by the hour or aggregated over 1 full day (24-hour period)

Send us an email at sales@iveda.com to ask about our free AI Software Program for Police Agencies

Key Features

  • Real-time alerts based on object type, face recognition, license plate recognition, abnormal events, intrusion detectionShow detected event on a map based on camera location
  • Live Tracking – Provides the exact movement and route of the target person/object based on its location and the detected time
  • Live View – Provides viewing of multiple live cameras on the same screen. View live scenes on the map with exact camera locations
  • GDPR – Allows faces and license plates to be blurred for privacy. Redaction time is 10x faster than manual methods
  • Administration and System Management – Camera management, Activate and deactivate AI functions on individual cameras. SMTP notification configuration

Sample Applications

Investigations

Help law enforcement agencies to quickly detect, identify and locate a person, vehicle , or object within seconds

Traffic Management

Classify and quantify vehicles and people, identify high-traffic time, space, and frequency to improve traffic flow

Campus Security

Heighten security with automated alerts to prevent known troublemakers from causing harm to school property, staff, and students

Loss Prevention

Provide evidence to immediately dispute or prove employee or customer claims for appropriate resolution

Highlights

  • Save time with high-accuracy, stable performance: obtain results within seconds (average search time: 0.2 seconds/query)
  • Maximize ease of use: simple setup, easy scale-up, easy integration with existing NVR/VMS or use as a stand-alone unit
  • Minimize manual effort: decrease man hour, decrease false positives to save time and effort (typical return on investment: 1.5 – 8 months)
  • Cost-competitive one-time purchase (no recurring fees) + optional upgrade & maintenance


Related Press Releases:

IvedaAI Pilot in Charter School and Banking and Law Enforcement in Mexico
Iveda Introduces Patented Next-Gen AI Search Technology for Video Surveillance