MDSC-Net: multi-directional spatial connectivity for road extraction in remote sensing images
Bibliographic record
Abstract
Extracting roads from complex remote sensing images is a crucial task for applications, such as autonomous driving, path planning, and road navigation. However, conventional convolutional neural network-based road extraction methods mostly rely on square convolutions or dilated convolutions in the local spatial domain. In multi-directional continuous road segmentation, these approaches can lead to poor road connectivity and non-smooth boundaries. Additionally, road areas occluded by shadows, buildings, and vegetation cannot be accurately predicted, which can also affect the connectivity of road segmentation and the smoothness of boundaries. To address these issues, this work proposes a multi-directional spatial connectivity network (MDSC-Net) based on multi-directional strip convolutions. Specifically, we first design a multi-directional spatial pyramid module that utilizes a multi-scale and multi-directional feature fusion to capture the connectivity relationships between neighborhood pixels, effectively distinguishing narrow and scale different roads, and improving the topological connectivity of the roads. Second, we construct an edge residual connection module to continuously learn and integrate the road boundaries and detailed information of shallow feature maps into deep feature maps, which is crucial for the smoothness of road boundaries. Additionally, we devise a high-low threshold connectivity algorithm to extract road pixels obscured by shadows, buildings, and vegetation, further refining textures and road details. Extensive experiments on two distinct public benchmarks, DeepGlobe and Ottawa datasets, demonstrate that MDSC-Net outperforms state-of-the-art methods in extracting road connectivity and boundary smoothness. The source code will be made publicly available at https://github/LYY199873/MDSC-Net.
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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| 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".