Experimental Settings of the Proposed Industrial Defect Detection Model

78.9% [email protected] was achieved by the proposed model on the NEU-DET benchmark for industrial defect detection. This dataset, shared by Jin Huang on figshare, contains experimental settings and performance metrics for a model designed to handle small samples and complex textures. It was last updated in May 2026.

Use Cases

• Benchmarking new defect detection algorithms based on reported mAP scores
• Studying model performance on small sample sizes and complex textures as described
• Comparing hierarchical attention mechanisms against baseline models like YOLOv8

Strengths

• Performance metrics ([email protected] scores of 78.9%, 84.7%, and 87.4%) are explicitly provided for three benchmark datasets
• The description details specific model components like a Semantic-Guided Query Enhancement Module

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download

Provenance

Source

Jin Huang

Collection Method

Likely contains experimental parameters and results from model evaluation on benchmark datasets.

Freshness

Last updated 2026-05-12 17:37:57