Pith. sign in

REVIEW 19 cited by

Can We Generate Images with CoT? Let's Verify and Reinforce Image Generation Step by Step

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2501.13926 v2 pith:BAXQJPSA submitted 2025-01-23 cs.CV cs.AIcs.CL

classification cs.CVcs.AIcs.CL
keywords generationimageautoregressivemodelsparmreasoningmodelpotential
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Chain-of-Thought (CoT) reasoning has been extensively explored in large models to tackle complex understanding tasks. However, it still remains an open question whether such strategies can be applied to verifying and reinforcing image generation scenarios. In this paper, we provide the first comprehensive investigation of the potential of CoT reasoning to enhance autoregressive image generation. We focus on three techniques: scaling test-time computation for verification, aligning model preferences with Direct Preference Optimization (DPO), and integrating these techniques for complementary effects. Our results demonstrate that these approaches can be effectively adapted and combined to significantly improve image generation performance. Furthermore, given the pivotal role of reward models in our findings, we propose the Potential Assessment Reward Model (PARM) and PARM++, specialized for autoregressive image generation. PARM adaptively assesses each generation step through a potential assessment approach, merging the strengths of existing reward models, and PARM++ further introduces a reflection mechanism to self-correct the generated unsatisfactory image, which is the first to incorporate reflection in autoregressive image generation. Using our investigated reasoning strategies, we enhance a baseline model, Show-o, to achieve superior results, with a significant +24% improvement on the GenEval benchmark, surpassing Stable Diffusion 3 by +15%. We hope our study provides unique insights and paves a new path for integrating CoT reasoning with autoregressive image generation. Code and models are released at https://github.com/ZiyuGuo99/Image-Generation-CoT

Discussion (0). Sign in to comment.

Forward citations

Cited by 19 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Demystifying Video Reasoning

    cs.CV 2026-03 conditional novelty 7.0 of 10

    Video diffusion models reason along the denoising trajectory (Chain-of-Steps), not primarily across frames, and this mechanism can be nudged by ensembling latent trajectories.

  2. MultiRef: Controllable Image Generation with Multiple Visual References

    cs.CV 2025-08 conditional novelty 7.0 of 10

    MultiRef-bench shows that current image generators that accept multiple visual references still fail to combine them reliably, with the best tested model OmniGen reaching only 66.6% synthetic and 79.0% real-world alig...

  3. FocusDiff: Advancing Fine-Grained Text-Image Alignment for Autoregressive Visual Generation through RL

    cs.CV 2025-06 conditional novelty 7.0 of 10

    FocusDiff improves autoregressive text-to-image generation by training on paired similar prompts with a modified GRPO objective, achieving state-of-the-art alignment on PairComp and gains on GenEval and T2I-CompBench.

  4. R2I-Bench: Benchmarking Reasoning-Driven Text-to-Image Generation

    cs.CV 2025-05 conditional novelty 7.0 of 10

    A 3,068-prompt benchmark with per-instance Q&A scoring shows that current text-to-image models, including reasoning-enhanced ones, handle reasoning-driven prompts poorly, with mathematical reasoning near zero.

  5. Model Guides You How to Draw: Adaptive Visual Gating for Unified Multimodal Reasoning

    cs.CV 2026-07 conditional novelty 6.5 of 10

    Early Generation Intent and Visual Fidelity signals let AdaViG abort low-utility intermediate image generations in UMM math CoT, improving accuracy up to 5.7% and cutting visual FLOPs 25–91%.

  6. Reconstruction Alignment Improves Unified Multimodal Models

    cs.CV 2025-09 conditional novelty 6.0 of 10

    RECA, a self-supervised post-training objective that conditions unified multimodal models on their own visual understanding embeddings to reconstruct input images, improves text-to-image and editing benchmarks across ...

  7. Echo-4o: Harnessing the Power of GPT-4o Synthetic Images for Improved Image Generation

    cs.CV 2025-08 conditional novelty 6.0 of 10

    A curated GPT-4o synthetic image dataset improves open-source generation models on instruction-following, surreal scenes, and multi-reference synthesis, plus two new benchmarks to measure those skills.

  8. Think-Before-Draw: Decomposing Emotion Semantics & Fine-Grained Controllable Expressive Talking Head Generation

    cs.CV 2025-07 conditional novelty 6.0 of 10

    A two-stage text-guidance framework, Think-Before-Draw, uses chain-of-thought prompting to convert emotion labels into facial muscle descriptions and then progressively guides a diffusion model from coarse emotion to ...

  9. ComplexBench-Edit: Benchmarking Complex Instruction-Driven Image Editing via Compositional Dependencies

    cs.CV 2025-06 conditional novelty 6.0 of 10

    Introduces a benchmark for chain-dependent image editing instructions plus a region-aware consistency metric, and shows a chain-of-thought prompt improves a Gemini-based editor.

  10. Q-Ponder: A Unified Training Pipeline for Reasoning-based Visual Quality Assessment

    cs.CV 2025-06 conditional novelty 6.0 of 10

    Q-Ponder is a two-stage pipeline (distill-then-reinforce) that makes a 7B multimodal model both more accurate at image quality scoring and better at explaining its judgments.

  11. TIIF-Bench: How Does Your T2I Model Follow Your Instructions?

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A new benchmark with 5,000 varied prompts and a VLM-based fine-grained evaluation protocol aims to measure and rank text-to-image models' instruction-following ability.

  12. RePrompt: Reasoning-Augmented Reprompting for Text-to-Image Generation via Reinforcement Learning

    cs.CV 2025-05 conditional novelty 6.0 of 10

    RePrompt uses RL-trained reasoning traces to enhance text-to-image prompts, boosting spatial composition and counting scores across FLUX, SD3, and PixArt-Σ while keeping image generators fixed.

  13. MINT-CoT: Enabling Interleaved Visual Tokens in Mathematical Chain-of-Thought Reasoning

    cs.CV 2025-06 conditional novelty 5.0 of 10

    MINT-CoT-7B interleaves fine-grained visual tokens into each math reasoning step and reports 73.70 on MathVista-Math, 64.72 on GeoQA, and 69.6 on MMStar-Math.

  14. Draw ALL Your Imagine: A Holistic Benchmark and Agent Framework for Complex Instruction-based Image Generation

    cs.CV 2025-05 conditional novelty 5.0 of 10

    A new benchmark and agent framework for complex text-to-image generation, with an unvalidated AI-judge evaluation and claims that the agent outperforms GPT-4o on the authors' own benchmark.

  15. VF-Eval: Evaluating Multimodal LLMs for Generating Feedback on AIGC Videos

    cs.CV 2025-05 conditional novelty 5.0 of 10

    A new benchmark, VF-Eval, measures how well multimodal LLMs check, detect, and reason about errors in AI-generated videos, and shows frontier models remain far below human performance.

  16. UniRL: Self-Improving Unified Multimodal Models via Supervised and Reinforcement Learning

    cs.CV 2025-05 conditional novelty 5.0 of 10

    A self-improving post-training method that uses a model's own generated images as training data, with SFT and GRPO, improves generation and understanding and reduces task imbalance.

  17. Self-Reflective Reinforcement Learning for Diffusion-based Image Reasoning Generation

    cs.CV 2025-05 reject novelty 5.0 of 10

    SRRL trains diffusion models to reason about images by repeatedly denoising, evaluating, and re-noising their own samples across several reflective rounds.

  18. ReasonGen-R1: CoT for Autoregressive Image generation models through SFT and RL

    cs.CV 2025-05 reject novelty 4.0 of 10

    ReasonGen-R1 reports that adding a GPT-generated text rationale before image generation, trained with SFT and GRPO, improves Janus-Pro-7B on GenEval, DPG-Bench, and T2I-Benchmark.

  19. Reinforcement Fine-Tuning Powers Reasoning Capability of Multimodal Large Language Models

    cs.CL 2025-05 conditional novelty 2.0 of 10

    A survey-style position paper claims that reinforcement fine-tuning powers reasoning in multimodal LLMs, summarizing over a hundred recent works and proposing five future research directions.

Pith tools