Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

The proposed industrial defect detection architecture integrates semantic guidance and hierarchical attention.

It combines a query enhancement mechanism and a multi-scale feature fusion module to improve detection accuracy and structural modeling capability.

More »

Fig 1 Expand

Fig 2.

The proposed framework of the query enhancement mechanism with semantic guidance aims to optimize the initial query representation through multi-source semantic fusion and residual feedback.

The semantic pathways are implemented by lightweight attention blocks operating on query tokens, and semantic priors are derived from multi-scale encoder features via pooling and linear projection within the same network. This design enhances the modeling capability of multi-scale defect features and improves the consistency of attention responses.

More »

Fig 2 Expand

Fig 3.

The overall architecture of the proposed Hierarchical Attention Fusion Structure integrates multi-scale convolutional encoding and an attention fusion mechanism.

The fused representation is delivered to the Feature Aggregation Buffer for scale aligned consolidation and then mapped by the Branch-wise Generator to form branch specific candidate features. This structure is designed to enhance the representation capability of defect features and to improve semantic consistency modeling.

More »

Fig 3 Expand

Fig 4.

Ten typical defects in NEU-DET.

More »

Fig 4 Expand

Fig 5.

DAGM2007 Ten typical defects of medium and high precision industrial surfaces.

More »

Fig 5 Expand

Fig 6.

Ten typical defects of printed circuit boards in PCB-DET.

More »

Fig 6 Expand

Table 1.

Experimental Settings of the Proposed Model.

More »

Table 1 Expand

Table 2.

Comparison of Defect Detection Results on NEU-DET Dataset.

More »

Table 2 Expand

Table 3.

Comparison of Defect Detection Results on DAGM2007 Dataset.

More »

Table 3 Expand

Table 4.

Comparison of Defect Detection Results on PCB-DET Dataset.

More »

Table 4 Expand

Fig 7.

The impact of the number of layers of the hierarchical attention module on the experimental results.

More »

Fig 7 Expand

Table 5.

Ablation study of the proposed modules on three datasets.

More »

Table 5 Expand

Fig 8.

Qualitative experimental results on the NEU-DET dataset.

More »

Fig 8 Expand

Fig 9.

Qualitative experimental results on the DAMG2007 dataset.

More »

Fig 9 Expand

Fig 10.

Qualitative experimental results on the PCB-DET dataset.

More »

Fig 10 Expand

Fig 11.

Grad-Cam experimental results on the PCB-DET dataset.

More »

Fig 11 Expand

Fig 12.

Grad-Cam experimental results on the NEU-GRAD dataset.

More »

Fig 12 Expand

Fig 13.

Reliability diagrams at 30% label noise on NEU-DET, DAGM2007, and PCB-DET.

Compared with the RT-DETR baseline, our QEM-SG + HAF method yields confidence–accuracy curves closer to the ideal diagonal, indicating better calibration under noisy annotations.

More »

Fig 13 Expand