A few-shot learning approach for the segmentation of subsurface defects in thermography images of concrete structures
Bibliographic record
Abstract
The identification and categorization of subsurface damages in thermal images of concrete structures remain an ongoing challenge that demands expert knowledge. Consequently, creating a substantial number of annotated samples for training deep neural networks poses a significant issue. Artificial intelligence (AI) models particularly encounter the problem of false positives arising from thermal patterns on concrete surfaces that do not correspond to subsurface damages. Such false detections would be easily identifiable in visible images, underscoring the advantage of possessing additional information about the sample surface through visible imaging. In light of these challenges, this study proposes an approach that employs a few-shot learning method known as the Siamese Neural Network (SNN), to frame the problem of subsurface delamination detection in concrete structures as a multi-modal similarity region comparison problem. The proposed procedure is evaluated using a dataset comprising 500 registered pairs of infrared and visible images captured in various infrastructure scenarios. Our findings indicate that leveraging prior knowledge regarding the similarity between visible and thermal data can significantly reduce the rate of false positive detection by AI models in thermal images.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".