3-D Line Matching Network Based on Matching Existence Guidance and Knowledge Distillation
Bibliographic record
Abstract
In applications, such as scene reconstruction and odometry, accurate matching associations for 3-D lines are crucial. Real-world scenes introduce inconsistencies due to variations in perspective, leading to nonoverlapping data acting as noise. Accurately matching partially overlapping sets of 3-D lines becomes challenging, potentially resulting in failed scene reconstruction and erroneous positioning. Prior approaches relied on the traditional iterative closest line (ICL) methods, involving iterative calculations and sensitivity to initial poses, and were prone to matching failures in low-overlap rate data and singular pattern scenes. Existing 3-D line matching networks either did not consider the noise in 3-D line collections or failed to retain more valid matching pairs, while these models often require a larger number of parameters and inference time. To address these issues, this article proposes matching existence guidance module (MEG)-Net, a Plücker line matching network guided by the existence of matches. It leverages the rich geometric characteristics of 3-D lines represented as Plücker lines, enhancing feature robustness. By guiding the model to handle the noisy data through match existence guidance, it improves the model’s performance on partially overlapping 3-D line data. Experiments on the indoor and outdoor data sets and the Out of Distribution (OOD) data sets demonstrate that the MEG-Net outperforms traditional methods and baseline models in 3-D line matching, with better scalability and noise robustness, achieving state-of-the-art results. Additionally, we propose an innovative knowledge distillation method based on the matching matrices, training a more efficient MEG-Net mini student model with approximately 70% fewer parameters and multiply accumulate operations (MACs), while maintaining superior performance and faster inference speeds on the indoor data sets.
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 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.005 | 0.002 |
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".