Synergistic Coordinate and Attention Module for Pipeline Weld Defect Detection
An enhanced YOLOv8n-based detection method achieves a Precision of 88.3% and an [email protected] of 79.2% on the ROC-DET dataset, improving industrial pipeline inner-wall defect inspection according to study data. By integrating a synergistic coordinate and attention module alongside a hybrid CIoU–NWD loss, the upgraded model boosts spatial localization and accurately identifies small, low-contrast defects that … Read more