CAOShurong中文

Publications

Publications

Research in wafer inspection and industrial anomaly detection.

FALCO-WAFER architecture, showing the backbone, encoder, and decoder
FALCO-WAFER

FALCO-WAFER: Feature-Aware Lightweight Contextual Detector for Wafer Defect Detection

Haotian Zhang*, Shurong Cao*, Ningmu Zou

IEEE International Test Conference in Asia (ITC-Asia), pp. 43–47, 2025.

A lightweight, feature-aware detector for semiconductor wafer inspection. Evaluated on 5,723 labeled defect images: 90.7% AP@0.5, 7.19% false-negative rate, and 13.3M parameters.

About this work

The architecture combines a Multi-Scale Depthwise Block for efficient texture encoding with a Token-Energy Diagonal Attention head for feature refinement. The work addresses subtle, low-contrast wafer defects while keeping the model compact for manufacturing inspection. My research work included data preparation, model benchmarking, architecture improvement, and scientific writing.

Cite this work
@inproceedings{zhang2025falcowafer,
  title={FALCO-WAFER: Feature-Aware Lightweight Contextual Detector for Wafer Defect Detection},
  author={Zhang, Haotian and Cao, Shurong and Zou, Ningmu},
  booktitle={2025 IEEE International Test Conference in Asia (ITC-Asia)},
  pages={43--47},
  year={2025},
  doi={10.1109/ITC-Asia67627.2025.00016}
}
Download .bib
DOI / IEEE ↗Code ↗
Texture-AD benchmark design: different product textures across training and testing
Texture-AD

Texture-AD: An Anomaly Detection Dataset and Benchmark for Real Algorithm Development

Tianwu Lei*, Bohan Wang*, Silin Chen, Shurong Cao, Ningmu Zou

arXiv:2409.06367, 2024 · CVM 2026 conference program.

An industrial anomaly-detection benchmark covering 15 cloth types, 14 semiconductor wafer types, and 10 metal-plate types, with pixel-level defect annotations.

About this work

Texture-AD examines the gap between algorithm development and production-line inspection through representative textures acquired under different optical schemes. The benchmark supports evaluation of how methods generalize across products.

Cite this work
@misc{lei2024texturead,
  title={Texture-AD: An Anomaly Detection Dataset and Benchmark for Real Algorithm Development},
  author={Lei, Tianwu and Wang, Bohan and Chen, Silin and Cao, Shurong and Zou, Ningmu},
  year={2024},
  eprint={2409.06367},
  archivePrefix={arXiv}
}
Download .bib
arXiv ↗PDF ↗