AlignVAR: Towards Globally Consistent Visual Autoregression for Image Super-Resolution
This work proposes a globally consistent framework for image super-resolution, offering a new paradigm for efficient high-quality reconstruction, which is incremental in advancing autoregressive methods for this domain.
The paper tackled the problem of applying visual autoregressive models to image super-resolution by addressing challenges like locality-biased attention and residual-only supervision, resulting in improved structural coherence and perceptual fidelity with over 10x faster inference and nearly 50% fewer parameters than diffusion-based methods.
Visual autoregressive (VAR) models have recently emerged as a promising alternative for image generation, offering stable training, non-iterative inference, and high-fidelity synthesis through next-scale prediction. This encourages the exploration of VAR for image super-resolution (ISR), yet its application remains underexplored and faces two critical challenges: locality-biased attention, which fragments spatial structures, and residual-only supervision, which accumulates errors across scales, severely compromises global consistency of reconstructed images. To address these issues, we propose AlignVAR, a globally consistent visual autoregressive framework tailored for ISR, featuring two key components: (1) Spatial Consistency Autoregression (SCA), which applies an adaptive mask to reweight attention toward structurally correlated regions, thereby mitigating excessive locality and enhancing long-range dependencies; and (2) Hierarchical Consistency Constraint (HCC), which augments residual learning with full reconstruction supervision at each scale, exposing accumulated deviations early and stabilizing the coarse-to-fine refinement process. Extensive experiments demonstrate that AlignVAR consistently enhances structural coherence and perceptual fidelity over existing generative methods, while delivering over 10x faster inference with nearly 50% fewer parameters than leading diffusion-based approaches, establishing a new paradigm for efficient ISR.