A CNN-Transformer Hybrid for Precise Object Detection in UAV Aerial Imagery
Bibliographic record
Abstract
The proposed hybrid framework in this paper is a deep learning technique to robustly detect objects and classify them in real-time applications and includes the use of Convolutional Neural Networks (CNNs) to capture local features, Transformer blocks to learn the global context, and adaptive feature fusing gating to integrate information to achieve optimal consistency. The proposed architecture uses feature pyramid networks (FPN) combined with path aggregation networks (PAN) to help provide better multi-scale representation and accelerate large and small, as well as rich and weak objects. Broad experiments have been carried out across several benchmark and cross-domain datasets such as indoor detection, agricultural leaf diseases recognition, and X-ray illegal content identification. This proposed model has an mAP@0.5:0.95 of 55.7 percent, a precision of 91-4 percent, and a recall of 88.0 percent, with a speed of 126 FPS on an RTX 4090, out beating state-of-the-art models like YOLOv7, DETR, and RT-DETR. Superior generalization ability is confirmed by cross-domain evaluations, and complementary role of the modules is validated by ablation studies. These findings show that the suggested framework can accomplish a desirable trade-off of accuracy and speed with good results and may be utilized on real-time tasks, e.g., autonomous navigation, industrial inspection, or security monitoring.
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 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| 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".