SEA-YOLO: Thermal Imaging Object Detection via Spatial Edge Attention for Low-visibility Autonomous Driving
Bibliographic record
Abstract
Autonomous driving demands safer and more reliable intelligent perception systems, particularly for object detection. Recent deep learning–based approaches usually rely on RGB (visual spectrum) images for object detection as a basis for environment perception. However, visual cameras perform poorly under low-visibility conditions, reducing their effectiveness in adverse environments. To address this issue, infrared (IR) thermal imaging has been introduced, as it can capture objects based on heat signatures that are often invisible in RGB. Nevertheless, thermal images also suffer from drawbacks, including low resolution, high noise, and unclear boundaries, which restrict their performance in complex, real-world scenarios. Thus, many IR-based object detection models combine RGB and IR modalities to improve robustness, though at the cost of increased architectural complexity and computational time. To overcome these challenges, this research proposes an enhanced single-modality thermal imaging–based convolutional neural network (CNN). The network integrates an efficient spatial edge extractor, coupled with an attention mechanism, into a YOLO backbone, for edge-aware learning, which we term the Spatial Edge Attention YOLO (SEA-YOLO). This integration enables the network to leverage objects’ boundary information more effectively, thereby improving object detection accuracy. Experimental results on multiple benchmark datasets demonstrate that the proposed approach achieves a strong detection rate, proving its potential for safer and more reliable driving assistance systems (DAS) in low-visibility conditions.
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 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 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".