A novel learning-based feature recognition method using multiple sectional view representation
- Submitting institution
-
The University of Huddersfield
- Unit of assessment
- 12 - Engineering
- Output identifier
- 11
- Type
- D - Journal article
- DOI
-
10.1007/s10845-020-01533-w
- Title of journal
- Journal of Intelligent Manufacturing
- Article number
- -
- First page
- 1291
- Volume
- 31
- Issue
- 5
- ISSN
- 0956-5515
- Open access status
- Compliant
- Month of publication
- January
- Year of publication
- 2020
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
-
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Smart manufacturing requires AI to interpret manufacturing semantics. A novel machine learning technique is presented, named MsvNet, used to detect machined features from a 3D model. This important work can recognise, locate single and multi-features, outperforms state-of-the-art feature recognition, greatly reducing training samples with improved recognition performance. The pioneer work and results attracted interest from world-leading CAD company Dassault Systems (Spatial, https://www.3ds.com). This work also initiated a multi-disciplinary collaboration with Machine Learning and Perception Lab (MLP@UoM), School of Computer Science, University of Manchester (Dr Ke Chen ke.chen@manchester.ac.uk). It has led to follow-on research as part of EPSRC project (EP/S001328/1).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
