RestoRect: Degraded Image Restoration via Latent Rectified Flow & Feature Distillation
This addresses the problem of slow high-performance models for image restoration, offering a practical solution for applications requiring efficient processing, though it appears incremental as it builds on existing distillation and flow techniques.
The paper tackles the trade-off between performance and speed in degraded image restoration by proposing RestoRect, a method that uses latent rectified flow and feature distillation to achieve faster inference and convergence while maintaining quality, demonstrating superior results across 15 datasets and 8 metrics.
Current approaches for restoration of degraded images face a critical trade-off: high-performance models are too slow for practical use, while fast models produce poor results. Knowledge distillation transfers teacher knowledge to students, but existing static feature matching methods cannot capture how modern transformer architectures dynamically generate features. We propose 'RestoRect', a novel Latent Rectified Flow Feature Distillation method for restoring degraded images. We apply rectified flow to reformulate feature distillation as a generative process where students learn to synthesize teacher-quality features through learnable trajectories in latent space. Our framework combines Retinex theory for physics-based decomposition with learnable anisotropic diffusion constraints, and trigonometric color space polarization. We introduce a Feature Layer Extraction loss for robust knowledge transfer between different network architectures through cross-normalized transformer feature alignment with percentile-based outlier detection. RestoRect achieves better training stability, and faster convergence and inference while preserving restoration quality. We demonstrate superior results across 15 image restoration datasets, covering 4 tasks, on 8 metrics.