Exploring Diffusion with Test-Time Training on Efficient Image Restoration
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917142012624896 |
|---|---|
| author | Lu, Rongchang Luo, Tianduo Jiang, Yunzhi Yue, Conghan Yang, Pei Liu, Guibao Gu, Changyang |
| author_facet | Lu, Rongchang Luo, Tianduo Jiang, Yunzhi Yue, Conghan Yang, Pei Liu, Guibao Gu, Changyang |
| contents | Image restoration faces challenges including ineffective feature fusion, computational bottlenecks and inefficient diffusion processes. To address these, we propose DiffRWKVIR, a novel framework unifying Test-Time Training (TTT) with efficient diffusion. Our approach introduces three key innovations: (1) Omni-Scale 2D State Evolution extends RWKV's location-dependent parameterization to hierarchical multi-directional 2D scanning, enabling global contextual awareness with linear complexity O(L); (2) Chunk-Optimized Flash Processing accelerates intra-chunk parallelism by 3.2x via contiguous chunk processing (O(LCd) complexity), reducing sequential dependencies and computational overhead; (3) Prior-Guided Efficient Diffusion extracts a compact Image Prior Representation (IPR) in only 5-20 steps, proving 45% faster training/inference than DiffIR while solving computational inefficiency in denoising. Evaluated across super-resolution and inpainting benchmarks (Set5, Set14, BSD100, Urban100, Places365), DiffRWKVIR outperforms SwinIR, HAT, and MambaIR/v2 in PSNR, SSIM, LPIPS, and efficiency metrics. Our method establishes a new paradigm for adaptive, high-efficiency image restoration with optimized hardware utilization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_14541 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Exploring Diffusion with Test-Time Training on Efficient Image Restoration Lu, Rongchang Luo, Tianduo Jiang, Yunzhi Yue, Conghan Yang, Pei Liu, Guibao Gu, Changyang Computer Vision and Pattern Recognition I.4.9 Image restoration faces challenges including ineffective feature fusion, computational bottlenecks and inefficient diffusion processes. To address these, we propose DiffRWKVIR, a novel framework unifying Test-Time Training (TTT) with efficient diffusion. Our approach introduces three key innovations: (1) Omni-Scale 2D State Evolution extends RWKV's location-dependent parameterization to hierarchical multi-directional 2D scanning, enabling global contextual awareness with linear complexity O(L); (2) Chunk-Optimized Flash Processing accelerates intra-chunk parallelism by 3.2x via contiguous chunk processing (O(LCd) complexity), reducing sequential dependencies and computational overhead; (3) Prior-Guided Efficient Diffusion extracts a compact Image Prior Representation (IPR) in only 5-20 steps, proving 45% faster training/inference than DiffIR while solving computational inefficiency in denoising. Evaluated across super-resolution and inpainting benchmarks (Set5, Set14, BSD100, Urban100, Places365), DiffRWKVIR outperforms SwinIR, HAT, and MambaIR/v2 in PSNR, SSIM, LPIPS, and efficiency metrics. Our method establishes a new paradigm for adaptive, high-efficiency image restoration with optimized hardware utilization. |
| title | Exploring Diffusion with Test-Time Training on Efficient Image Restoration |
| topic | Computer Vision and Pattern Recognition I.4.9 |
| url | https://arxiv.org/abs/2506.14541 |