V-Co: A Closer Look at Visual Representation Alignment via Co-Denoising

• URL: http://arxiv.org/abs/2603.16792v1
• Date: Tue, 17 Mar 2026 17:01:54 GMT
• Title: V-Co: A Closer Look at Visual Representation Alignment via Co-Denoising
• Abstract summary: We present V-Co, a systematic study of visual co-denoising in a unified JiT-based framework.<n>Our study reveals four key ingredients for effective visual co-denoising.<n>V-Co outperforms the underlying pixel-space diffusion baseline and strong prior pixel-diffusion methods.
• Score: 65.5867130156805
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Pixel-space diffusion has recently re-emerged as a strong alternative to latent diffusion, enabling high-quality generation without pretrained autoencoders. However, standard pixel-space diffusion models receive relatively weak semantic supervision and are not explicitly designed to capture high-level visual structure. Recent representation-alignment methods (e.g., REPA) suggest that pretrained visual features can substantially improve diffusion training, and visual co-denoising has emerged as a promising direction for incorporating such features into the generative process. However, existing co-denoising approaches often entangle multiple design choices, making it unclear which design choices are truly essential. Therefore, we present V-Co, a systematic study of visual co-denoising in a unified JiT-based framework. This controlled setting allows us to isolate the ingredients that make visual co-denoising effective. Our study reveals four key ingredients for effective visual co-denoising. First, preserving feature-specific computation while enabling flexible cross-stream interaction motivates a fully dual-stream architecture. Second, effective classifier-free guidance (CFG) requires a structurally defined unconditional prediction. Third, stronger semantic supervision is best provided by a perceptual-drifting hybrid loss. Fourth, stable co-denoising further requires proper cross-stream calibration, which we realize through RMS-based feature rescaling. Together, these findings yield a simple recipe for visual co-denoising. Experiments on ImageNet-256 show that, at comparable model sizes, V-Co outperforms the underlying pixel-space diffusion baseline and strong prior pixel-diffusion methods while using fewer training epochs, offering practical guidance for future representation-aligned generative models.

Related papers

• Where Does Generative Difficulty Reside? An Empirical Study of Target Representations [62.54876287800644]
  The target representation defines the distribution an image generator must learn, yet it is often treated as an interchangeable interface.<n>We study raw pixels, SD-VAE latents and DINOv2 as well as MAE representation-autoencoder features within a unified masked autoregressive rectified-flow model.<n>Our results show that compression, reconstruction fidelity, token dimensionality, and visible semantic clustering do not individually predict generative behavior.
  arXiv  Detail & Related papers  (2026-08-01T12:32:31Z)
• Anchoring and Steering Diffusion: Enhancing the Faithfulness of Text-to-Image Generation at Inference Time [16.04233421396159]
  AnchorSteer is a training-free framework that exerts fine-grained control over.<n>textbfAnchorSteer and its components.<n>Experiments on GenEval and T2I-CompBench++ demonstrate that AnchorSteer consistently outperforms existing baselines in text--image alignment.
  arXiv  Detail & Related papers  (2026-07-29T09:08:57Z)
• Improving Visual Representation Alignment Generation with GRPO [51.071351994330605]
  diffusion transformers have demonstrated strong image synthesis capabilities but remain inefficient to train due to weak alignment.<n>We propose VRPO, a reinforcement-based optimization strategy that replaces REPA's static alignment loss with a generative representation policy optimization objective.<n>Our VRPO-driven training seamlessly integrates into diffusion transformers, introducing negligible cost and preserving full compatibility with SiT and DiT architectures.
  arXiv  Detail & Related papers  (2026-05-30T07:21:40Z)
• Guiding Diffusion-based Reconstruction with Contrastive Signals for Balanced Visual Representation [81.40978077888693]
  Contrastive Language-Image Pre-training (CLIP) has become a key bottleneck for downstream performance.<n>Recent solutions use diffusion models to enhance representations by conditioning image reconstruction on CLIP visual tokens.<n>We integrate contrastive signals into diffusion-based reconstruction to pursue more comprehensive visual representations.
  arXiv  Detail & Related papers  (2026-03-05T04:45:49Z)
• Prior-guided Hierarchical Instance-pixel Contrastive Learning for Ultrasound Speckle Noise Suppression [2.7777929779304955]
  We propose a prior-guided hierarchical instance-pixel contrastive learning model for ultrasound denoising.<n>A statistics-guided pixel-level contrastive learning strategy is introduced to enhance distributional discrepancies between noisy and clean pixels.<n>A hybrid Transformer-CNN architecture is adopted, coupling a Transformer-based encoder for global context modeling with a CNN-based decoder optimized for fine-grained anatomical structure restoration.
  arXiv  Detail & Related papers  (2026-02-14T16:01:58Z)
• Edit2Perceive: Image Editing Diffusion Models Are Strong Dense Perceivers [55.15722080205737]
  Edit2Perceive is a unified diffusion framework that adapts editing models for depth, normal, and matting.<n>Our single-step deterministic inference yields up to faster runtime while training on relatively small datasets.
  arXiv  Detail & Related papers  (2025-11-24T01:13:51Z)
• Prototype-Guided Diffusion: Visual Conditioning without External Memory [2.1155908599769764]
  Prototype Diffusion Model integrates prototype learning directly into the diffusion process for efficient visual conditioning - without external memory.<n>PDM maintains high generation quality while reducing computational and storage overhead, offering a scalable alternative to retrieval-based conditioning in diffusion models.
  arXiv  Detail & Related papers  (2025-08-13T16:18:35Z)
• "Principal Components" Enable A New Language of Images [79.45806370905775]
  We introduce a novel visual tokenization framework that embeds a provable PCA-like structure into the latent token space.<n>Our approach achieves state-of-the-art reconstruction performance and enables better interpretability to align with the human vision system.
  arXiv  Detail & Related papers  (2025-03-11T17:59:41Z)
• One-Step Diffusion Model for Image Motion-Deblurring [85.76149042561507]
  We propose a one-step diffusion model for deblurring (OSDD), a novel framework that reduces the denoising process to a single step.<n>To tackle fidelity loss in diffusion models, we introduce an enhanced variational autoencoder (eVAE), which improves structural restoration.<n>Our method achieves strong performance on both full and no-reference metrics.
  arXiv  Detail & Related papers  (2025-03-09T09:39:57Z)
• Diffusion-Based Representation Learning [65.55681678004038]
  We augment the denoising score matching framework to enable representation learning without any supervised signal. In contrast, the introduced diffusion-based representation learning relies on a new formulation of the denoising score matching objective. Using the same approach, we propose to learn an infinite-dimensional latent code that achieves improvements of state-of-the-art models on semi-supervised image classification.
  arXiv  Detail & Related papers  (2021-05-29T09:26:02Z)
• Generalized Octave Convolutions for Learned Multi-Frequency Image Compression [20.504561050200365]
  We propose the first learned multi-frequency image compression and entropy coding approach. It is based on the recently developed octave convolutions to factorize the latents into high and low frequency (resolution) components. We show that the proposed generalized octave convolution can improve the performance of other auto-encoder-based computer vision tasks.
  arXiv  Detail & Related papers  (2020-02-24T01:35:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.