FiRe: Fixed-Noise Refinement for Visual Counterfactual Explanations

• URL: http://arxiv.org/abs/2608.08664v1
• Date: Sun, 09 Aug 2026 12:15:37 GMT
• Title: FiRe: Fixed-Noise Refinement for Visual Counterfactual Explanations
• Abstract summary: We propose FiRe, a fixed-noise Refinement framework for visual counterfactual explanations.<n>FiRe maps the input to a fixed noise level and iteratively refines the noisy state at that level.<n>FiRe achieves about 3$times$ faster online inference and 8$times$ fewer FLOPs while obtaining comparable or state-of-the-art counterfactual quality.
• Score: 9.955401099290096
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Visual counterfactual explanations aim to change classifier decisions through realistic and localized edits while preserving decision-irrelevant content. Existing DDPM-based methods typically perform classifier-guided editing along a long reverse denoising trajectory. The changing noise levels make semantic editability and spatial control difficult to balance, and the editable state is noisy, whereas the target classifier is trained on clean images. As a result, these methods require either costly recursive denoising or low-quality one-step estimates to obtain classifier-facing clean images. We propose FiRe, a Fixed-noise Refinement framework for visual counterfactual explanations. Rather than following a reverse denoising trajectory, FiRe maps the input to a fixed noise level and iteratively refines the noisy state at that level. To provide clean images for classifier guidance, FiRe first adapts Pixel Mean Flow to visual counterfactual explanation, enabling direct clean-image prediction from noisy states. To make fixed-noise refinement produce minimal and localized counterfactual edits, FiRe introduces three FiRe-specific controls: a dynamic dual-mask strategy, adaptive guidance, and early stopping, which determine where edits accumulate, which changes become visible, and when refinement stops. Experiments on five tasks across three datasets show that, compared with the strongest recent baseline, FiRe achieves about 3$\times$ faster online inference and 8$\times$ fewer FLOPs while obtaining comparable or state-of-the-art counterfactual quality.

Related papers

• RIDGE: Re-Noising with Internal Dynamic Guidance for Image Editing [11.37302017486576]
  Inversion-free flow-based image editing avoids latent inversion, but still requires a target-side state at every editing step.<n>Ridge is an inversion-free and training-free method that maintains the edited state as an evolving approximation to the unavailable clean target state.<n> Experiments on two benchmarks using two backbones, SD3 Medium and FLUX.1-dev, show that RIDGE offers a favorable aggregate trade-off among source preservation, target alignment, and perceptual quality.
  arXiv  Detail & Related papers  (2026-08-04T03:20:09Z)
• Adaptive Inference-Time Scaling via Early-Step Latent Verification for Image Editing [49.40982704974727]
  An unsuitable initial noise may lead to unsatisfactory editing results.<n>Recent inference-time scaling methods address this issue by sampling multiple initial noises and selecting better candidates.<n>We propose VeriLatent, a plug-and-play adaptive inference-time scaling framework with early-step latent verification for image editing.
  arXiv  Detail & Related papers  (2026-06-13T08:30:33Z)
• DuET: Dual Expert Trajectories for Diffusion Image Editing [42.49842620609682]
  We introduce DuET, a training-free inference method that temporarily relaxes source-image conditioning.<n>DuET consistently improves instruction relevance, semantic fidelity, and perceptual quality across diverse models and benchmarks.
  arXiv  Detail & Related papers  (2026-06-11T12:58:48Z)
• DirectEdit: Step-Level Accurate Inversion for Flow-Based Image Editing [51.56484100374058]
  We propose DirectEdit, a training-free editing method for pre-trained text-to-image (T2I) models.<n>DirectEdit eliminates the inherent reconstruction error without introducing additional neural function evaluations (NFEs)<n>Experiments demonstrate that DirectEdit achieves efficient and accurate image editing, delivering superior performance that outperforms state-of-the-art methods.
  arXiv  Detail & Related papers  (2026-05-04T10:09:18Z)
• Beyond Fixed Inference: Quantitative Flow Matching for Adaptive Image Denoising [5.719237048831079]
  Diffusion and flow-based generative models have shown strong potential for image restoration.<n>Image denoising under unknown and varying noise conditions remains challenging.<n>We propose a quantitative flow matching framework for adaptive image denoising.
  arXiv  Detail & Related papers  (2026-04-02T08:29:54Z)
• Detection-Driven Object Count Optimization for Text-to-Image Diffusion Models [54.641726517633025]
  We propose a new framework that uses pre-trained object counting techniques and object detectors to guide generation.<n>First, we optimize a counting token using an outer-loop loss computed on fully generated images.<n>Second, we introduce a detection-driven scaling term that corrects errors caused by viewpoint and proportion shifts.
  arXiv  Detail & Related papers  (2024-08-21T15:51:46Z)
• TurboEdit: Text-Based Image Editing Using Few-Step Diffusion Models [53.757752110493215]
  We focus on a popular line of text-based editing frameworks - the edit-friendly'' DDPM-noise inversion approach. We analyze its application to fast sampling methods and categorize its failures into two classes: the appearance of visual artifacts, and insufficient editing strength. We propose a pseudo-guidance approach that efficiently increases the magnitude of edits without introducing new artifacts.
  arXiv  Detail & Related papers  (2024-08-01T17:27:28Z)
• Denoising as Adaptation: Noise-Space Domain Adaptation for Image Restoration [64.84134880709625]
  We show that it is possible to perform domain adaptation via the noise space using diffusion models.<n>In particular, by leveraging the unique property of how auxiliary conditional inputs influence the multi-step denoising process, we derive a meaningful diffusion loss.<n>We present crucial strategies such as channel-shuffling layer and residual-swapping contrastive learning in the diffusion model.
  arXiv  Detail & Related papers  (2024-06-26T17:40:30Z)
• Drag Your Noise: Interactive Point-based Editing via Diffusion Semantic Propagation [30.737586652869457]
  DragNoise offers robust and accelerated editing without retracing the latent map. The bottleneck features of U-Net inherently possess semantically rich features ideal for interactive editing. DragNoise achieves superior control and semantic retention, reducing the optimization time by over 50% compared to DragDiffusion.
  arXiv  Detail & Related papers  (2024-04-01T11:09:40Z)
• NLIP: Noise-robust Language-Image Pre-training [95.13287735264937]
  We propose a principled Noise-robust Language-Image Pre-training framework (NLIP) to stabilize pre-training via two schemes: noise-harmonization and noise-completion. Our NLIP can alleviate the common noise effects during image-text pre-training in a more efficient way.
  arXiv  Detail & Related papers  (2022-12-14T08:19:30Z)
• Dual Adversarial Network: Toward Real-world Noise Removal and Noise Generation [52.75909685172843]
  Real-world image noise removal is a long-standing yet very challenging task in computer vision. We propose a novel unified framework to deal with the noise removal and noise generation tasks. Our method learns the joint distribution of the clean-noisy image pairs.
  arXiv  Detail & Related papers  (2020-07-12T09:16:06Z)

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.