Nexplant Visual Insight

AI-powered intelligent visual data analytics platform specially designed for factory-generated image data analytics

Deploying a complex, sophisticated AI process is just one click away

Expert-level techniques and knowledge are brought together from factory data analytics experts to create and provide an integrated, easy-to-use AI Ops platform for non-expert users. Nexplant VI helps you run real-time inspection to predict defective products and thereby improve productivity. Equally notable is it immediately detects dangerous behaviors of factory workers in production lines and senses defects in equipment and facilities in advance to ensure the factory site continues to be run in a safe environment.
  • Accuracy rate to extract product faulty 98% on average
  • Leakage rate of faulty products 330ppm or less
  • Extract rate compared to that with naked eyes 3 times more Secured

01

04

Benefits

Meet AI-based video analysis intelligence platform specialized in manufacturing environment

    • Equipment Management for Hyper-connectivity
    • Nexplant VI easily transfers and shares data between cameras (vision, infrared thermal, CCTV, etc.), edge appliance and the visual analysis platform, and its equipment management feature helps you view the whole set of data at one glance.
      Once model training is completed, the model can easily be deployed into edge appliance through UI. This significantly reduces the workload of system administrators.
    • Optimized Infrastructure for Real-time, High-speed Operation & Analysis
    • GPU farm-based resource optimization minimizes the necessary server resources and enables real-time, high-speed computation of complicated operations. Its offering set of Nexplant VI with HW, servers and edge appliances will help you more flexibly respond to different site scales and environments. Infrastructure custom-built for an individual customer is also available.
    • High Performance, Accurate Analysis at Minimum Cost
    • Nexplant VI reduces the time to develop analysis algorithms to the minimum level and empowers non-expert users to succeed in high performance modeling. It also helps the analysis system continuously maintain its operational reliability and accuracy.

Features

  • Easy AI Model Lifecycle Management

    Non-expert users can develop and build AI-powered models easily, and maintain or further improve the model performance continuously with the easy-to-run AI Ops platform.

  • Refinement Features for Convenient Modeling

    Enjoy Nexplant VI’s outstanding features – Visual Search, Auto Tagging, Image Refinement and Dataset Configuration while spending significantly less time on running the features and enhancing the model’s performance.

  • Special Features for Best Model Selection

    Nexplant VI self-finds an optimal model most suitable for the data collected and makes model recommendations. It also features performance comparison of the trained models, making complicated modeling process simpler and easier.

  • Use of Edge Devices

    Apply the edge computing technology for real-time, high-speed computation. This helps you spend less the network deployment cost and address latency issues for growing field data volumes.

  • Auto-analysis of Product Defects

    Auto-analysis of high-resolution images helps you detect product defects and find root causes. You can get the improvement of the overall productivity by managing and controlling the process ahead of time.

  • Safety Management for Equipment and Facilities

    Monitoring CCTV cameras automatically detect abnormal running of factory equipment and facilities in real time to proactively prevent safety incidents and keep the workplace safe.

  • Safety Management for Workers

    Nexplant VI senses potentially dangerous situations that could arise at the workplace and sends instant alerts to ensure the safety of workers in real time. This will prevent workplace accidents and injuries and help stakeholders take immediate action.

Platform Architecture

Nexplant Visual Insight Platform Architecture

AI Business Service
  • Intelligent Inspection - Defect Detection
  • Intelligent Safety Monitoring - Incident Prevention, Violation Detection
  • Design Optimization - Pattern Search, Drawing Review
  • Predictive Maintenance - Predictive Maintenance, Anomaly Classification
AI Service Module
  • AI algorithm - Machine Learning, Deep Learning
  • AI Operation Process - (Virtous Cycle) Collect, Clean, Train, Deploy, Decect/Classify, Review
AI Infrastructure
  • Computing Resource - High-Performance GPU Farm Computing, Edge Device
  • Data Resource - Product Design Data, Monitoring Data, Quality Data

Use Cases

    Specifications

    Recommended specifications

    • HW(WEB/WAS)

      – CPU: 2.1Ghz, 8cores or above
      – Memory: 32GB or above
      – Storage: 2TB or above

    • HW(DB/Storage)

      – CPU: 2.1Ghz, 4 cores or above
      – Memory: 32GB or above
      – Storage: 5TB or above

    • HW (Training Server)

      – CPU: 2.1Ghz, 8 cores or above
      – Memory: 128GB or above
      – GPU: V100 x 2
      (※ If distributed, v100 x 4, Node (HW) 2 units (x2) or above)
      – Storage: 2TB or above

    • HW (Server for Proprietary Features)

      (※ If Auto Tagging, BMP, Simulation, Image Search are included: HW x 2, If the Detection feature is added, HW x 3)
      – CPU: 2.1Ghz, 8 cores or above
      – Memory: 64GB or above
      – GPU: V100 x 2
      – Storage: 2TB or above

    • HW (Decision Server)

      – CPU: 2.1Ghz, 8 cores or above
      – Memory: 64GB or above
      – GPU: V100 x 1
      – Storage: 1TB or above

    • SW

      – (WEB/WAS) Server OS: Linux-64bit
      – User Agent Browser: Internet Explorer 11 or above

    • Let’s talk
    • Whether you’re looking for a specific business solution or just need some questions answered, we’re here to help