WearFormer-MO: Code, Model Snapshot and Evaluation Cells for TBM Disc-Cutter Wear Vision Measurement

Fuente: Zenodo
Saved in:
Bibliographic Details
Main Authors: Liu, Guangyuan, Hongmei, Wang, Runqi, Li, Huiping, Zhang, Hongfei, Xie, Bright, Liu, Siyong, Zhang, Guizhen, Liu, Hongtao, Zhu, Dun, Liu, Sai, Ma
Format: Recurso digital
Published: Zenodo 2027
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866902103949049856
author Liu, Guangyuan
Hongmei, Wang
Runqi, Li
Huiping, Zhang
Hongfei, Xie
Bright, Liu
Siyong, Zhang
Guizhen, Liu
Hongtao, Zhu
Dun, Liu
Sai, Ma
author_facet Liu, Guangyuan
Hongmei, Wang
Runqi, Li
Huiping, Zhang
Hongfei, Xie
Bright, Liu
Siyong, Zhang
Guizhen, Liu
Hongtao, Zhu
Dun, Liu
Sai, Ma
contents <p>Code, frozen model snapshot, and full evaluation outputs (1500 per-config cells, aggregated metrics, figure data, cached SWAG ViT-B/16 features) accompanying the manuscript 'Vision Measurement of TBM Disc-Cutter Wear Using Reliability-Weighted Multi-View Aggregation and Ordinal-Metric Estimation' submitted to Measurement Science and Technology (2026). The image dataset (5.2 GB) will be added in a subsequent version of this record once the industrial co-author institutions complete their public-release authorisation. Headline numbers (8-view inference, 5-seed mean): ICS protocol MAE 0.33 mm and R² 0.987; LOCO protocol on unseen cutters MAE 2.44 mm and R² 0.80.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_20396792
institution Zenodo
language
publishDate 2027
publisher Zenodo
record_format zenodo
spellingShingle WearFormer-MO: Code, Model Snapshot and Evaluation Cells for TBM Disc-Cutter Wear Vision Measurement
Liu, Guangyuan
Hongmei, Wang
Runqi, Li
Huiping, Zhang
Hongfei, Xie
Bright, Liu
Siyong, Zhang
Guizhen, Liu
Hongtao, Zhu
Dun, Liu
Sai, Ma
"Tunnel boring machine", "TBM", "Disc cutter wear", "Machine vision", "Multi-view learning", "Reliability-weighted attention", "Ordinal regression", "Vision Transformer", "SWAG pretraining", "Industrial measurement", "Underground construction"
<p>Code, frozen model snapshot, and full evaluation outputs (1500 per-config cells, aggregated metrics, figure data, cached SWAG ViT-B/16 features) accompanying the manuscript 'Vision Measurement of TBM Disc-Cutter Wear Using Reliability-Weighted Multi-View Aggregation and Ordinal-Metric Estimation' submitted to Measurement Science and Technology (2026). The image dataset (5.2 GB) will be added in a subsequent version of this record once the industrial co-author institutions complete their public-release authorisation. Headline numbers (8-view inference, 5-seed mean): ICS protocol MAE 0.33 mm and R² 0.987; LOCO protocol on unseen cutters MAE 2.44 mm and R² 0.80.</p>
title WearFormer-MO: Code, Model Snapshot and Evaluation Cells for TBM Disc-Cutter Wear Vision Measurement
topic "Tunnel boring machine", "TBM", "Disc cutter wear", "Machine vision", "Multi-view learning", "Reliability-weighted attention", "Ordinal regression", "Vision Transformer", "SWAG pretraining", "Industrial measurement", "Underground construction"
url https://doi.org/10.5281/zenodo.20396792