logo

close

Company

Product

Tech & Solution

Support

  • image
  • image

AIBOX AIBOX-B404B

  • Easy to configure and intuitive AI video analytics applications
  • Easy AI system construction with existing CCTV cameras
  • Multi level AI based false alarm reduction
  • High performance pose estimation technology
  • Accurate event detection through scene understanding
  • Multiple AI apps can be run on each video channel
  • Easy integration into other systems
  • Web based configuration
  • Enhanced security options
  • Combining rules with other events
  • Supports up to 16 channels maximum.
  • NDAA Compliant
Smart CCTV surveillance with deep learning-based embedded AI BOX

AI BOX provides analyzed images and alarms to VMS and NVR by deep learning analysis of CCTV images.
It provides various solutions throughout more than 40 types of AI apps such as behavioral pattern analysis, store, safety, intrusion management,
vehicle control and fire management and so on and two or more complex rules can be applied simultaneously to one channel.
Watch AI BOX demo video

AI BOX
Key Features
4-channel deep learning-based embedded AI BOX
01 02 03
Embedded AI device Simultaneous
multi-algorithm processing 		Advanced false alarm
reduction filter
04 05 06
Minimize false alarm
by min & max size filter 		Easy addition to
existing CCTV system 		AI NPU + Image DSP
compound CPU
07 08 09
RTSP AI video re-streaming Case suitable for
industrial environment
(-30℃~70℃) 		Low power consumption
(6Watts / DC12V)

AI BOX
Edge Computing AI

Minimal server burden and real-time signal processing by Edge device
Low latency compared to central signal processing method, less risk of hacking, reduce cost and power consumption
Edge device for minimizing server load, real-time signal processing, reduced latency, and cost savings
Traffic overload and signal delay in server-based video analysis

AI BOX
Advanced false alarm filter

Advanced false alarm filter
AI BOX is applied with an Advanced multi-stage false alarm filter that filters the factors that cause the false alarm to prevent it from repeating when the false alarm occurs repeatedly due to shaking, reflection, or changing the tracking ID with an adjacent object
AI BOX false alarm filter for event notification filtering and accuracy improvemen

Others : Re-recognized object occurs repeated alarm

AI BOX : Re-recognized object doesn't occur repeated alarm

AI BOX
3rd Generation Intelligent CCTV

3rd generation Hybrid intelligent technology
Even with the top-ranked deep-learning based object detection Co-DETR, the recognition rate is only around 66%, AI deep-learning alone can not reduce misrecognition. With 3rd generation technology, AI BOX can provide accurate detection than other AI solution.
　- 1st gen : DSP (Digital Signal Processor)
　- 2nd gen : NPU (Neural Processing Unit)
　- 3rd gen : Hybrid Intelligent Technology that Combines Deep Learning AI with Video Analysis to process DSP and NPU simultaneously
Hybrid intelligent technology combining deep learning AI and VA with DSP and NPU processing

AI BOX
AI Application list
Basic features of deep learning-based AI BOX
Basic and advanced features of deep learning-based AI BOX
Extra licenses for deep learning-based AI BOX

AI BOX
AI application demo video

Traffic control
It provides various vehicle control images and information using LPR, vehicle counting and stop detection as a combination rule
AI-based object recognition for apartment entrance vehicle entry and exit counting
Hybrid technology for vehicle type, number, manufacturer, and LPR number recognition
Traffic monitoring center for lane-specific traffic volume and crosswalk intrusion vehicle detection
Lane-specific vehicle counting and crosswalk pedestrian monitoring system

Analysis of human behavior
It provides a various deep learning images and information using apps such as crowd, fall down, imminent threat, people count and loitering detection apps
Time-based congestion analysis and frequency detection with congestion situation object counter alarm
Real-time fall detection and rapid incident notification to minimize human casualties
Threat detection using behavior recognition and scene understanding technology
Advanced people counting technology for building and store entry/exit headcount
Detection of wandering people, vehicles, and motorcycles for suspicious behavior detection and enhanced security

Privacy protection
This video is an example of providing various images and information by applying two or more apps such as privacy masking and area intrusion detection to single channel in a complex rule

Dynamic personal data masking excluding event-triggering objects
Real-time video masking technology for protecting personal data of unidentified individuals

Thermal intrusion camera
It provides human detection images and information by analyzing images from thermal camera used for special purpose such as military units

Safety management
It provides images and information of the situation where safety helmets and masks are not worn at industrial sites by deep learning analyze
Forklift operator helmet and safety gear non-wear detection technology

Fire detection
It provides images and information of the fire situation by deep learning.
This app will be more effective together with fire detection camera (NK2100BL-AF-F2) which has built-in fire detection sensor
AI video analysis technology for rapid smoke detection to prevent major fires
Rapid smoke and fire detection to prevent major disasters

Object tracking
It provides auto track and zoom in on moving object using PTZ track app

Hand & Foot intrusion and Intentional body gaze detector
It provides images and information by deep learning situation where person continues to observe in a specific direction or invade hands or feet
Multi-alarm notification for detecting specific direction and hand/foot intrusion
Deep learning analysis for detecting specific direction observation and hand/foot intrusion with video and data

3rd party VMS integration
This is an example of integrating images and metadata of AI BOX with 3rd party's VMS
LIST
위로