analiza aiqbench
This paper introduces CART, a novel framework that dynamically allocates computational resources and task-specific expertise to handle multiple co-occurring image degradations simultaneously. By routing patches based on local restoration difficulty and conditioning on global task features, it overcomes the limitations of uniform computation in existing all-in-one methods. The approach achieves state-of-the-art results on composite degradation benchmarks, marking a significant step forward in efficient computer vision models.
Learning Where and What to Restore for Composite Image Restoration
arXiv:2503.17915v2 Announce Type: replace-cross Abstract: Real-world degraded images often contain multiple co-occurring degradation types, making composite image restoration fundamentally different from the single-degradation setting assumed by most existing all-in-one methods. These methods typically apply uniform spatial computation and single-label task conditioning, limiting both spatial adaptivity and explicit modeling of degradation mixtures. We propose CART (Composite-Adaptive Routing and Task Conditioning), a unified framework that jointly decides where to spend computation and what degradation cues should guide restoration: a spatial mixer routes patches to graded-capacity experts by local restoration difficulty, and a channel mixer routes each image through degradation-specific experts, conditioned on the global task feature with multi-hot classification supervision. Empirically, CART achieves state-of-the-art performance for composite degradation restoration on CDD-11 and delivers the best results to date on the conventional 3-task and 5-task all-in-one benchmarks.
Abstract:Real-world degraded images often contain multiple co-occurring degradation types, making composite image restoration fundamentally different from the single-degradation setting assumed by most existing all-in-one methods. These methods typically apply uniform spatial computation and single-label task conditioning, limiting both spatial adaptivity and explicit modeling of degradation mixtures. We propose CART (Composite-Adaptive Routing and Task Conditioning), a unified framework that jointly decides where to spend computation and what degradation cues should guide restoration: a spatial mixer routes patches to graded-capacity experts by local restoration difficulty, and a channel mixer routes each image through degradation-specific experts, conditioned on the global task feature with multi-hot classification supervision. Empirically, CART achieves state-of-the-art performance for composite degradation restoration on CDD-11 and delivers the best results to date on the conventional 3-task and 5-task all-in-one benchmarks.
Accepted to NeurIPS 2026. 16 pages, 7 figures, 4 tables. v2: substantially revised camera-ready version; title changed from "Cat-AIR: Content and Task-Aware All-in-One Image Restoration"; method renamed from Cat-AIR to CART and extended to composite degradations
Image and Video Processing (eess.IV); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2503.17915 [eess.IV]
(or arXiv:2503.17915v2 [eess.IV] for this version)
Text integral sub licența deschisă a publicației · 190 cuvinte
Subiecte conexe
Exploring self-distilled reasoning for supervised fine-tuning with Amazon Nova
AWS · 21 iul. 2026
From Rollout to Reset: A Graph-Based Harness for Autonomous Long-Horizon Manipulation Evaluation
arXiv · 18 sept. 2026
X-Restormer++: 1st Place Solution for the UG2+ CVPR 2026 All-Weather Restoration Challenge
arXiv · 3 iun. 2026
IntBMoE: Integrating Block-Level Conditioning into Expert Composition for Full-Participation Mixture-of-Experts
arXiv · 21 sept. 2026