DDR-Net: Dual-Stream for Degraded Infrared Image Restoration Network
Bibliographic record
Abstract
Infrared (IR) imaging technology is drawing significant attention in various critical applications due to its unique capabilities for object detection and scene understanding, even under challenging environments and adverse weather conditions. However, IR imaging fundamentally faces two major limitations: inherent lowresolution and contamination by complex noise (Gaussian, stripe, and non-uniformity). Existing approaches suffer from critical limitations. Conventional Super-Resolution (SR) models are vulnerable to composite noise in IR imagery. Existing Denoising (DN) models either address only single noise types or neglect the low-resolution problem. Sequential pipelines (DN→SR or SR→DN) introduce irreversible information loss, while current joint DN+SR methods consider only simplistic noise models, proving insufficient for real-world IR data. To address these challenges, this study proposes the Dual-stream for Degraded infrared image Restoration Network (DDR-Net), a lightweight end-to-end network for simultaneous complex noise removal and super-resolution. DDR-Net features three key contributions: (1) A novel frequency-domain decomposition-based dual-stream architecture that independently performs noise suppression in low-frequency components and edge preservation in high-frequency components, followed by efficient fusion. (2) The Self-Dual Calibrated Projection Attention (SDCPA) mechanism for effective information exchange between streams. (3) Integration of a validated composite noise model reflecting real-world IR sensor characteristics. Extensive experiments on four benchmark datasets (FLIR, IR100, DLS-NUC-100, and ESPOL FIR) demonstrate that DDR-Net consistently achieves superior performance across various noise levels and upscaling factors (×2, ×4). The parameter-efficient architecture ensures practical deployment suitability, providing an effective solution for infrared image restoration in diverse real-world scenarios.
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 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 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".