MBGN: multimodal BiFPN gate network for in-process GMAW defect detection

In-process defect monitoring is critical for guaranteeing the integrity of Gas Metal Arc Welding (GMAW) joints. Conventional non-destructive testing (NDT) delivers high-quality diagnostics only after the weld has finished, precluding immediate corrective action. We present the Multi-modal BiFPN Gate Network (MBGN), a deep-learning framework that fuses two complementary data streams: (i) time-synchronized welding current and voltage waveforms processed by a one-dimensional CNN, and (ii) high-speed molten-pool imagery processed by a ResNet backbone. The two modalities are merged through a Bidirectional Feature Pyramid Network (BiFPN), and a lightweight Gate module adaptively recalibrates cross-modal interactions before a final classifier. Experiments on an industrially collected multimodal dataset demonstrate that MBGN achieved 0.748 accuracy and 0.719 macro-F1 on the held-out evaluation set. Compared with the strongest task-matched multimodal baseline, MWFN, MBGN improved macro-F1 by 0.120 while maintaining comparable accuracy. These results indicate the potential of multimodal fusion for in-process GMAW defect monitoring under the investigated industrial conditions, while broader deployment still requires validation across additional materials, joint types, and online system configurations.