Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

BiRQA: Bidirectional Robust Quality Assessment for Images

Aleksandr Gushchin, Dmitriy S. Vatolin, Anastasia Antsiferova

TL;DR

BiRQA is the only FR IQA model combining competitive accuracy with real-time throughput and strong adversarial resilience, and to the authors' knowledge is the only FR IQA model combining competitive accuracy with real-time throughput and strong adversarial resilience.

Abstract

Full-Reference image quality assessment (FR IQA) is important for image compression, restoration and generative modeling, yet current neural metrics remain slow and vulnerable to adversarial perturbations. We present BiRQA, a compact FR IQA metric model that processes four fast complementary features within a bidirectional multiscale pyramid. A bottom-up attention module injects fine-scale cues into coarse levels through an uncertainty-aware gate, while a top-down cross-gating block routes semantic context back to high resolution. To enhance robustness, we introduce Anchored Adversarial Training, a theoretically grounded strategy that uses clean "anchor" samples and a ranking loss to bound pointwise prediction error under attacks. On five public FR IQA benchmarks BiRQA outperforms or matches the previous state of the art (SOTA) while running ~3x faster than previous SOTA models. Under unseen white-box attacks it lifts SROCC from 0.30-0.57 to 0.60-0.84 on KADID-10k, demonstrating substantial robustness gains. To our knowledge, BiRQA is the only FR IQA model combining competitive accuracy with real-time throughput and strong adversarial resilience.

BiRQA: Bidirectional Robust Quality Assessment for Images

TL;DR

BiRQA is the only FR IQA model combining competitive accuracy with real-time throughput and strong adversarial resilience, and to the authors' knowledge is the only FR IQA model combining competitive accuracy with real-time throughput and strong adversarial resilience.

Abstract

Full-Reference image quality assessment (FR IQA) is important for image compression, restoration and generative modeling, yet current neural metrics remain slow and vulnerable to adversarial perturbations. We present BiRQA, a compact FR IQA metric model that processes four fast complementary features within a bidirectional multiscale pyramid. A bottom-up attention module injects fine-scale cues into coarse levels through an uncertainty-aware gate, while a top-down cross-gating block routes semantic context back to high resolution. To enhance robustness, we introduce Anchored Adversarial Training, a theoretically grounded strategy that uses clean "anchor" samples and a ranking loss to bound pointwise prediction error under attacks. On five public FR IQA benchmarks BiRQA outperforms or matches the previous state of the art (SOTA) while running ~3x faster than previous SOTA models. Under unseen white-box attacks it lifts SROCC from 0.30-0.57 to 0.60-0.84 on KADID-10k, demonstrating substantial robustness gains. To our knowledge, BiRQA is the only FR IQA model combining competitive accuracy with real-time throughput and strong adversarial resilience.
Paper Structure (27 sections, 1 theorem, 20 equations, 6 figures, 10 tables, 1 algorithm)

This paper contains 27 sections, 1 theorem, 20 equations, 6 figures, 10 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Feature Extraction
  5. BiRQA Network
  6. Anchored Adversarial Training
  7. Pointwise Error Bound via Anchored Ranking Loss
  8. Assumptions.
  9. Proof sketch.
  10. Notes on practicality of the assumptions.
  11. Implementation Details
  12. Evaluation setup
  13. Results
  14. Conclusion
  15. Proof of Theorem 1
  16. ...and 12 more sections

Key Result

Theorem 1.1

Assume: Then Moreover, if the worst-error index $j^\star\in\arg\max_j|\tilde{y}_j-y_j|$ happens to be an anchor ($j^\star\in\mathcal{S}$), then $E\le \varepsilon$.

Figures (6)

  • Figure 1: Overall scheme of BiRQA. A reference–distorted pair yields four feature maps per pyramid level. Cross-Scale Residual Attention Module (CSRAM) and Spatial Cross-Gating Block (SCGB) allow the model to pass information in both directions between scales. A Reliability-Aware Head (GeM + dual MLPs) estimates per-level impact and reliability.
  • Figure 2: Scheme of the Cross-Scale Residual Attention Module (CSRAM) that lifts high-resolution cues to the next scale and injects them via uncertainty-aware gated residuals (strength $\alpha$, confidence $\rho$, and residual $R$) to refine the lower-resolution features.
  • Figure 3: Computational efficiency (FPS) vs. Performance (PLCC) comparison on PDAP-HDDS dataset with image size of $1920 \times 1080$ pixels. Our model achieves comparable PLCC to SOTA method TOPIQ, while being $\sim3$ times faster and having a notably smaller number of parameters.
  • Figure 4: SROCC and inference FPS on KADID-10k for different feature sets. The chosen quartet lies on the Pareto frontier, offering the best accuracy. FPS was measured on images with $1920\times1080$ resolution.
  • Figure 5: (a): Convergence of the Anchored Ranking Loss over 1,000 iterations. (b): Comparison of bounds, provided by Theorem 1 with empirical values of maximum pointwise errors.
  • ...and 1 more figures

Theorems & Definitions (2)

  • Theorem 1.1: Pointwise $\ell_\infty$ bound from max-near anchored hinge
  • proof