Semantic Image Inversion and Editing using Rectified Stochastic Differential Equations
read the original abstract
Generative models transform random noise into images; their inversion aims to transform images back to structured noise for recovery and editing. This paper addresses two key tasks: (i) inversion and (ii) editing of a real image using stochastic equivalents of rectified flow models (such as Flux). Although Diffusion Models (DMs) have recently dominated the field of generative modeling for images, their inversion presents faithfulness and editability challenges due to nonlinearities in drift and diffusion. Existing state-of-the-art DM inversion approaches rely on training of additional parameters or test-time optimization of latent variables; both are expensive in practice. Rectified Flows (RFs) offer a promising alternative to diffusion models, yet their inversion has been underexplored. We propose RF inversion using dynamic optimal control derived via a linear quadratic regulator. We prove that the resulting vector field is equivalent to a rectified stochastic differential equation. Additionally, we extend our framework to design a stochastic sampler for Flux. Our inversion method allows for state-of-the-art performance in zero-shot inversion and editing, outperforming prior works in stroke-to-image synthesis and semantic image editing, with large-scale human evaluations confirming user preference.
This paper has not been read by Pith yet.
Forward citations
Cited by 26 Pith papers
-
Improving Robotic Generalist Policies via Flow Reversal Steering
Flow Reversal Steering steers flow matching generalist policies by reversing suboptimal actions to nearby better modes, enabling improved zero-shot control, quick distillation, and RL bootstrapping in robotic manipulation.
-
Consistent-Inversion: Reverse Consistency Guidance for Structure-Preserving Visual Editing
Consistent-Inversion introduces reverse consistency guidance that corrects early target denoising steps by checking reversibility toward the source inversion trajectory under the original prompt.
-
DirectEdit: Step-Level Accurate Inversion for Flow-Based Image Editing
DirectEdit achieves step-level accurate inversion for flow-based image editing by directly aligning forward paths, using attention feature injection and mask-guided noise blending to balance fidelity and editability w...
-
UniGeo: Unifying Geometric Guidance for Camera-Controllable Image Editing via Video Models
UniGeo unifies geometric guidance across three levels in video models to reduce geometric drift and improve consistency in camera-controllable image editing.
-
UniEditBench: A Unified and Cost-Effective Benchmark for Image and Video Editing via Distilled MLLMs
UniEditBench unifies image and video editing evaluation with a nine-plus-eight operation taxonomy and cost-effective 4B/8B distilled MLLM evaluators that align with human judgments.
-
Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes
Text-to-3D models lose prompt sensitivity for out-of-distribution shapes due to sink traps but retain geometric diversity via unconditional priors, enabling a decoupled inversion method for robust editing.
-
Efficient Zero-Shot Inpainting with Decoupled Diffusion Guidance
A new decoupled diffusion guidance method enables efficient zero-shot inpainting by avoiding backpropagation through the denoiser while maintaining observation consistency and quality.
-
Exploring Cross-Modal Flows for Few-Shot Learning
FMA introduces flow matching for multi-step cross-modal feature alignment in few-shot learning, using fixed coupling, noise augmentation, and early-stopping to outperform one-step PEFT methods.
-
UniEdit-Flow: Unleashing Inversion and Editing in the Era of Flow Models
UniEdit-Flow presents tuning-free Uni-Inv and Uni-Edit methods for inversion and editing in flow models that achieve accurate reconstruction and robust region-preserving edits across generative models.
-
Root-Selecting Fixed-Point Inversion for Rectified Flows via Trajectory Straightness
SelFix selects straighter-trajectory fixed-point solutions for rectified flow inversion to improve real-image reconstruction and source-preserving editing.
-
scCBGM: Interpretable Single-Cell Counterfactual Editing
scCBGM adapts concept bottleneck generative models with skip connections and cross-covariance penalties for single-cell data, enabling interpretable counterfactual editing and showing superior combinatorial generaliza...
-
StreamEdit: Training-Free Video Editing via Few-Step Streaming Video Generation
StreamEdit enables high-quality training-free video editing by adapting streaming video generation models with dual-branch fast sampling, self-attention bridge, cross-attention grounding, source-oriented guidance, and...
-
StreamEdit: Training-Free Video Editing via Few-Step Streaming Video Generation
StreamGVE enables high-quality training-free video editing by converting the task to noise-to-data streaming generation with dual-branch fast sampling, self-attention bridges, cross-attention grounding, source-oriente...
-
VAGS: Velocity Adaptive Guidance Scale for Image Editing and Generation
VAGS adapts the CFG scale at each ODE step using velocity alignment signals to raise structural fidelity in editing and sample quality in generation over fixed-scale baselines.
-
StyleTextGen: Style-Conditioned Multilingual Scene Text Generation
StyleTextGen proposes a dual-branch style encoder, text style consistency loss, and mask-guided inference to achieve superior style consistency and cross-lingual performance in multilingual scene text generation on a ...
-
Revealing the Gap in Human and VLM Scene Perception through Counterfactual Semantic Saliency
VLMs exhibit size, center, and saliency biases in scene understanding, relying less on people than humans do, with size bias as a key driver of divergence.
-
LimeCross: Context-Conditioned Layered Image Editing with Structural Consistency
LimeCross enables text-guided editing of individual layers in composite images by conditioning on cross-layer context via bi-stream attention while preserving layer integrity and introducing the LayerEditBench benchmark.
-
FluSplat: Sparse-View 3D Editing without Test-Time Optimization
FluSplat trains a model with geometric alignment constraints on multi-view edits to produce consistent 3D scene edits from sparse views in a single forward pass without test-time optimization.
-
UniGeo: Unifying Geometric Guidance for Camera-Controllable Image Editing via Video Models
UniGeo adds unified geometric guidance at three levels in video models to reduce geometric drift and improve structural fidelity in camera-controllable image editing.
-
RL-RIG: A Generative Spatial Reasoner via Intrinsic Reflection
RL-RIG uses a generate-reflect-edit loop with reinforcement learning to improve spatial accuracy in image generation, reporting up to 11% gains over prior open-source models on scene-graph metrics.
-
DirectEdit: Step-Level Accurate Inversion for Flow-Based Image Editing
DirectEdit eliminates reconstruction error in flow-based image editing by aligning forward paths and applying attention feature injection with mask-guided noise blending.
-
KD-CVG: A Knowledge-Driven Approach for Creative Video Generation
KD-CVG uses an Advertising Creative Knowledge Base plus Semantic-Aware Retrieval and Multimodal Knowledge Reference modules to improve semantic alignment and motion realism in text-to-video generation for advertising.
-
UniGeo: Unifying Geometric Guidance for Camera-Controllable Image Editing via Video Models
UniGeo improves camera-controllable image editing by injecting point cloud geometry into a video diffusion model at the representation, architecture, and loss levels, achieving state-of-the-art geometric consistency o...
-
Encoder-Decoder Manifold Alignment for Idempotent Generation
Encoder-decoder manifold alignment framework to achieve exact idempotency in generative models.
-
RIVET: Robust Idempotent Voice Attribute Editing
RIVET enforces an idempotency objective during training of voice attribute editing models to improve robustness to noisy labels, outperforming standard training on controlled noise and the GLOBE dataset.
-
On the Controllability-Fidelity Frontier in Diffusion Editing
A study deriving mathematical formulations and bounds for diffusion editing objectives while empirically comparing methods on fidelity and control metrics and discussing ethical issues.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.