Engineer — AI Defect Detection for Metal Inner Wall (data annotation and labeling for defect classification)
Image
ClassificationLed an AI defect detection project for metal inner wall inspection by building and using labeled image datasets to support automated identification and categorization of defects. Performed defect feature labeling after collecting and organizing amplified camera images, then validated model performance to guide iterative improvements. Worked on data cleaning and repeated evaluation cycles to improve detection accuracy and stability for the target defect categories. • Collected amplified images of metal inner walls from cameras and organized the dataset • Annotated defect features/labels on inspection images for downstream model training • Conducted data cleaning to improve label quality and dataset consistency • Verified model effects and iteratively optimized the detection pipeline