This study investigates the use of Laser and Magnetic Flux Leakage (MFL) pipelinedata to develop a deep learning model for accurate detection and segmentation ofpipeline defects. Laser images are used to precisely identify defect regions andprovide labels for training a Mask R-CNN model for localizing and segmentingdefects in MFL signals. Unlike conventional datasets where ground-truth labelsare pixel-wise accurate, our labels are derived from a different sensor modality,resulting in misalignment and feature discrepancies between the laser and MFLdata. These discrepancies lead to label noise and domain shift. Our experimentsshow that training advanced object detection and segmentation models using onlylaser-derived labels does not achieve accurate defect localization in MFL signals.This underscores the need for models capable of handling label discrepancies andadapting across domains to ensure robust and scalable performance in real-worldpipeline defect detection.