Proposed AI-based automated pipelines for human error detection in construction. (IMAGE)
Caption
The proposed pipeline achieves greater than 90% accuracy in extracting information from structural design documents and in experiments identified 253 human errors, including 29 critical errors, demonstrating its potential for reducing manual checking burden.
Credit
Associate Professor Jaewook Jeong from Seoul National University of Science and Technology
Usage Restrictions
Cannot be reused without permission
License
Original content