A Hierarchical Feature Fusion and Dynamic Collaboration Framework for Robust Small Target Detection
Bibliographic record
Abstract
Small target detection is an important research direction in computer vision, widely applied in scenarios such as drone monitoring, remote sensing image analysis, and autonomous driving. However, as small targets occupy fewer pixels, contain limited feature information, and often appear in complex backgrounds, existing detection algorithms face shortcomings in accuracy and robustness. To address this, this paper proposes a novel small target detection algorithm that integrates hierarchical feature fusion with a spatial dynamic collaboration mechanism. The hierarchical feature fusion module (HFA) effectively combines shallow detail features with deep semantic features, greatly enhancing the feature representation capability for small targets. Meanwhile, the dynamic collaboration mechanism (DCCA) dynamically adjusts feature fusion weights and detection strategies based on target scale and density distribution, thereby further improving detection accuracy and robustness. Extensive experiments are conducted on datasets such as VisDrone, TinyPerson, and NWPU VHR-10. Results demonstrate that, compared to state-of-the-art models like YOLOv8 and YOLOv10, the proposed algorithm achieves significant improvements in precision, recall, and mAP, with mAP increasing by 2.1% to 3.2% and mAP-95 by 1.2% to 1.8%. Ablation studies further validate the complementarity of HFA and DCCA in optimizing model performance, confirming the algorithm’s superiority and robustness in complex scenarios. This research provides a novel technical route for small target detection and offers valuable references for practical applications.
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 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".