Comparative Study of YOLO Architectures for Automated Wood Defect Detection
Bibliographic record
Abstract
The North American residential construction industry relies on wood as the principal material for structural framing, as well as for kitchen cabinets, decorative trim moldings, and door casings. In this regard, wood quality is a key determinant of structural integrity and aesthetics in construction. Traditional wood defect inspection during construction and furniture manufacture is time-consuming, inconsistent, and error prone. Machine-vision technology could solve these issues and improve wood quality assessment. The use of automated defect detection systems can improve inspection efficiency and accuracy while reducing manual labour. This paper evaluates the performance of four advanced You Only Look Once (YOLO) object detection models: YOLOv5l-seg, YOLOv7-E6E, YOLOv8l, and YOLOv9e for automated wood defect identification. Each variant involves a balance between accuracy and computational efficiency. YOLOv5l-seg supports segmentation, YOLOv7-E6E improves feature extraction, YOLOv8l speeds up inference, and YOLOv9e uses transformer-based components for better detection. Using a dataset of 3,300 annotated images spanning ten defect types, it was found that YOLOv9e achieves the highest precision (90.15%), demonstrating strong potential for real-time wood inspection in construction and manufacturing workflows. The results are discussed in the context of their applicability to off-site construction systems for quality tracking and defect traceability.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".