Multitask Learning for Unified Optical Surface Inspection: A Single Shared-Encoder Architecture for Thermal Reconstruction, Phase Unwrapping, and Defect Detection

Key Points

• The aim is to develop a unified optical surface inspection framework utilizing multitask learning for improved thermal reconstruction, phase unwrapping, and defect detection.
• Implemented a single shared-encoder architecture for multitask learning
• Conducted optical surface inspections using thermal and phase data
• Analyzed performance across multiple tasks: reconstruction, unwrapping, and detection
• Demonstrated improved accuracy in thermal reconstruction and defect detection
• Achieved successful phase unwrapping leveraging shared features from the encoder
• Results indicate efficiency gains compared to traditional methods in optical inspections