This project implements automated defect detection in semiconductor inspection images using a combination of:
- FAST-MCD (Fast Minimum Covariance Determinant) for robust statistical outlier detection,
- Image structure classification (linear, flat, patterned), and
- Morphological post-processing
It reproduces and extends the approach described in
"Defect Inspection in Semiconductor Images Using FAST-MCD Method and Neural Network"
Lee, Kim et al., IEEE Access 2020.