Novel architecture based on dual‐view fusion for underground hidden distresses detection
Bibliographic record
Abstract
The detection of underground hidden distresses aims to locate the anomaly areas beneath the surface. This task faces significant challenges, including the scarcity of distress samples and the complex three-dimensional hidden structures of the distresses. Traditional approaches typically employ generative adversarial networks and manually designed object detectors to address these issues. Nonetheless, the current methods face challenges in maintaining semantic consistency between the generated samples and the real-world samples, and they can only detect anomalies based on features from a single image. To overcome these limitations, this paper proposes an innovative detection architecture that significantly enhances the performance of multi-object hidden distress detection by introducing a dual-view (horizontal and longitudinal) correlated generation model and a dual-stream detection mechanism. The approach offers a comprehensive subsurface analysis: The horizontal view captures large-scale anomalies, while the longitudinal view reveals vertical structural details, boosting multi-object distress detection. Specifically, a ground-penetrating radar image diffusion model (GPRDiff) is proposed to generate hidden distress images with dual-view correlation. Furthermore, this study designs a novel dual-view cross-information fusion transformer to achieve efficient fusion of dual-stream information. Experimental results demonstrate that using a combination of GPRDiff-generated images and real images as input, along with a joint-view guided dual-stream detector, significantly improves the detection accuracy of multi-object hidden distresses. This research not only fills the technical gap in multi-object generation within the field of 3D radar road detection but also provides new research insights and technical pathways for other detection industries.
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 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| 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".