Deep learning-based video analytics solution

  Convenient learning and flexible recognition

  High accuracy and easy data expansion

VLAD (Vision Learning for Advanced Detection) is a software solution for AI image analysis based on deep learning that can be applied to various fields such as semiconductor processes that require high reliability, automotive parts that go through different production processes for each part, as well as displays, electronics, batteries, food, security, and medicine.

• Training Tool
  - Create and manage learning models
  - Image labeling and management
  - Training Area Setup
  - Classify Training and validation data
  
• Report Viewer
  - Filter and view learning results
  - View a summary of learning results (recognition rate for each label)
  - View a detailed list of Evaluation results for each Image
  - Detailed view of learning result labeling images
  
• Testing Tool
  - Start/Stop test
  - Import and save Training files
  - Camera and learning model Settings
  - Camera view Modes
  - Viewing test summary results
  - Viewing detailed test results
  

• Classification
  
  
  - Recognizes the lion's share of a single image and classifies it according to the class you define
  - Supports Multi-label Classification
  - Applies to good/bad inspection, inspection image type classification
  
  
• Object Detection
  
  
  - Detects the type and location of each Object In the form of a rectangular box by classifying the characteristics and class of each Object when multiple objects exist In one image.
  - Multi Object detection model to detect multiple objects
  - Applied to pattern matching, Object location detection and counting
  
  
• Segmentation
  
  
  - Object detection displays the shape of the detected Object In the form of a pixel-accurate area.
  - Objects of the same class are divided by the same color
  - Applied to the inspection of microscopic defects (defects, damage, stains) and atypical types of defects
  
  
• Anomaly Detection
  
  
  - Unsupervised learning algorithm learns one type of Image without labeling and classifies images outside of the learned type as False.
  - When you need to divide into one class and other classes
  - When there are too many categories of classes and It is difficult to Collect data clearly.