• Goal

- Development of a deep learning model that can detect specific motion patterns of equipment through DVS, RGB, and ToF sensors

  • - Image segmentation, object detection model development
  • - Development of a module that can extract physical values from detected patterns
  • - Using OpenCV

- Facility-related data analysis model development

  • - Detect malfunctions of equipment through physical numerical data of equipment and input sensor data
  • - Development of anomaly detection model through basic regression model and deep learning model

- Model lightweight

  • - Application of model weight reduction method using quantization and knowledge distillation
  • - Presenting a model capable of fast and stable inference even in low power conditions