Fig. 10
Confusion matrices on GC10-DET. MSEOD-DDFusionNet (b) outperforms YOLOv11s (a) with higher diagonal purity, better recognition of Wf and Cg, and effective detection of the challenging defect Rp.
Fig. 10 | Scientific Reports
Confusion matrices on GC10-DET. MSEOD-DDFusionNet (b) outperforms YOLOv11s (a) with higher diagonal purity, better recognition of Wf and Cg, and effective detection of the challenging defect Rp.
