Research Overview

My research focuses on semantic segmentation and industrial AI systems, particularly under limited labeled data scenarios.

Key Directions

1. Semi-Supervised and Weakly-Supervised Segmentation

  • Cross pseudo supervision
  • Weak/strong augmentation consistency
  • Feature-level alignment
  • Uncertaininty estimation
  • Pseudo-label refinement
  • Protype-based learning

2. Industrial Machine Vision

  • Steel defect detection
  • Industrial pipe inspection
  • Bridge and civil structure visual inspection and condition assesment
  • Road and pavement assesment and inspection
  • Robust real-world deployment

Broad Research Interest

  • Intelligent/smart manufacturing solutions
  • Intelligent equipment and robotics
  • Machine vision and deep learning applications in industry
  • Semi-supervised and unsupervised deep learning methods
  • Vision-Language Foundation Models
  • LLM and agentic modeling for industrial solutions
  • Industrial robotic perception and visual control
  • Automated industrial inspection, anomaly detection, pattern recognition, and fault diagnosis
  • Digital manufacturing, human-machine collaboration
  • Intelligent machine fault diagnosis and prediction
  • Intelligent transportations, Vision based guidance (AGVs)
  • Medical image segmentation and analysis
  • Computational optimization and system analysis