{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T03:32:04Z","timestamp":1777865524837,"version":"3.51.4"},"reference-count":68,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100023542","name":"Centre for Perceptual and Interactive Intelligence (CPII) Ltd","doi-asserted-by":"publisher","award":["2023B1515130008,XW"],"award-info":[{"award-number":["2023B1515130008,XW"]}],"id":[{"id":"10.13039\/501100023542","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.01422","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"15329-15339","source":"Crossref","is-referenced-by-count":1,"title":["From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning"],"prefix":"10.1109","author":[{"given":"Le","family":"Zhuo","sequence":"first","affiliation":[{"name":"CUHK MMLab"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Liangbing","family":"Zhao","sequence":"additional","affiliation":[{"name":"KAUST"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sayak","family":"Paul","sequence":"additional","affiliation":[{"name":"Hugging Face"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yue","family":"Liao","sequence":"additional","affiliation":[{"name":"CUHK MMLab"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Renrui","family":"Zhang","sequence":"additional","affiliation":[{"name":"CUHK MMLab"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yi","family":"Xin","sequence":"additional","affiliation":[{"name":"Shanghai AI Lab"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Peng","family":"Gao","sequence":"additional","affiliation":[{"name":"Shanghai AI Lab"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mohamed","family":"Elhoseiny","sequence":"additional","affiliation":[{"name":"KAUST"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hongsheng","family":"Li","sequence":"additional","affiliation":[{"name":"CUHK MMLab"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","author":"Abdelrahman","year":"2023","journal-title":"Toddlerdiffusion: Interactive structured image generation with cascaded schr\u201d odinger bridge"},{"key":"ref2","author":"Ahn","year":"2024","journal-title":"A noise is worth diffusion guidance"},{"key":"ref3","author":"Bai","year":"2024","journal-title":"Zigzag diffusion sampling: The path to success is zigzag"},{"key":"ref4","author":"Bai","year":"2025","journal-title":"Qwen2.5-vl technical report"},{"key":"ref5","volume-title":"Improving image generation with better captions. Computer Science","author":"Betker","year":"2023"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.2307\/2334029"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01764"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72670-5_12"},{"key":"ref9","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-031-73411-3_5","article-title":"Pixart- \u03c3: Weak-to-strong training of diffusion transformer for 4 k text-to-image generation","volume-title":"Proceedings of the European Conference on Computer Vision (ECCV)","author":"Chen"},{"key":"ref10","author":"Chen","year":"2025","journal-title":"Januspro: Unified multimodal understanding and generation with data and model scaling"},{"key":"ref11","year":"2025","journal-title":"Deepseek-r1: Incentivizing reasoning capability in 11 ms via reinforcement learning"},{"key":"ref12","article-title":"Scaling rectified flow transformers for high-resolution image synthesis","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Esser"},{"key":"ref13","author":"Feng","year":"2023","journal-title":"Alphazero-like tree-search can guide large language model decoding and training"},{"key":"ref14","article-title":"Lumina-t2x: Scalable flow-based large diffusion transformer for flexible resolution generation","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Gao"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2270"},{"key":"ref16","author":"Guo","year":"2025","journal-title":"Can we generate images with cot? let\u2019s verify and reinforce image generation step by step"},{"key":"ref17","article-title":"Glore: When, where, and how to improve 11 m reasoning via global and local refinements","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Havrilla"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.595"},{"key":"ref19","article-title":"Denoising diffusion probabilistic models","author":"Ho","year":"2020","journal-title":"Advances in Neural Information Processing Systems (NeurIPS)"},{"key":"ref20","article-title":"Lora: Low-rank adaptation of large language models","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Hu"},{"key":"ref21","author":"Hurst","year":"2024","journal-title":"Gpt-4o system card"},{"key":"ref22","author":"Jiang","year":"2025","journal-title":"T2i-r1: Reinforcing image generation with collaborative semantic-level and token-level cot"},{"key":"ref23","author":"Kaplan","year":"2020","journal-title":"Scaling laws for neural language models"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1723"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1594"},{"key":"ref26","author":"Kumar","year":"2024","journal-title":"Training language models to self-correct via reinforcement learning"},{"key":"ref27","volume-title":"Flux","year":"2024"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51701.2025.01453"},{"key":"ref29","author":"Li","year":"2024","journal-title":"Hunyuan-dit: A powerful multi-resolution diffusion transformer with fine-grained chinese understanding"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51701.2025.01763"},{"key":"ref31","article-title":"Flow matching for generative modeling","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Lipman"},{"key":"ref32","author":"Liu","year":"2024","journal-title":"Playground v3: Improving text-to-image alignment with deep-fusion large language models"},{"key":"ref33","author":"Liu","year":"2025","journal-title":"Improving video generation with human feedback"},{"key":"ref34","article-title":"Flow straight and fast: Learning to generate and transfer data with rectified flow","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Liu"},{"key":"ref35","author":"Ma","year":"2025","journal-title":"Inference-time scaling for diffusion models beyond scaling denoising steps"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2019"},{"key":"ref37","first-page":"35779","article-title":"Prodigy: an expeditiously adaptive parameter-free learner","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Mishchenko"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00387"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0970"},{"key":"ref40","volume-title":"Sdxl: Improving latent diffusion models for high-resolution image synthesis","author":"Podell","year":"2023"},{"key":"ref41","author":"Qi","year":"2024","journal-title":"Not all noises are created equally: Diffusion noise selection and optimization"},{"key":"ref42","author":"Qin","year":"2025","journal-title":"Luminaimage 2.0: A unified and efficient image generative framework"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1754"},{"key":"ref44","author":"Ren","year":"2024","journal-title":"Grounded sam: Assembling open-world models for diverse visual tasks"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"ref46","author":"Segalis","year":"2023","journal-title":"A picture is worth a thousand words: Principled recaptioning improves image generation"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0377"},{"key":"ref48","author":"Singhal","year":"2025","journal-title":"A general framework for inference-time scaling and steering of diffusion models"},{"key":"ref49","author":"Snell","year":"2024","journal-title":"Scaling 11 m test-time compute optimally can be more effective than scaling model parameters"},{"key":"ref50","article-title":"Denoising diffusion implicit models","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Song"},{"key":"ref51","article-title":"Score-based generative modeling through stochastic differential equations","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Song"},{"key":"ref52","author":"Tan","year":"2024","journal-title":"Ominicontrol: Minimal and universal control for diffusion transformer"},{"key":"ref53","author":"Team","year":"2023","journal-title":"Gemini: a family of highly capable multimodal models"},{"key":"ref54","volume-title":"Lumina-image 2.0: A unified and efficient image generative model","year":"2025"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00786"},{"key":"ref56","article-title":"Omniedit: Building image editing generalist models through specialist supervision","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Wei"},{"key":"ref57","article-title":"Generating sequences by learning to self-correct","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Welleck"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1109\/iccv51070.2023.00200"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.01241"},{"key":"ref60","article-title":"Sana: Efficient high-resolution image synthesis with linear diffusion transformers","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Xie"},{"key":"ref61","author":"Xie","year":"2025","journal-title":"Sana 1.5: Efficient scaling of training-time and inference-time compute in linear diffusion transformer"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0700"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0517"},{"key":"ref64","author":"Ye","year":"2023","journal-title":"Ipadapter: Text compatible image prompt adapter for text-toimage diffusion models"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0704"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00355"},{"key":"ref67","author":"Zhou","year":"2024","journal-title":"Golden noise for diffusion models: A learning framework"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.52202\/079017-4172"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11444455.pdf?arnumber=11444455","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T06:46:17Z","timestamp":1777531577000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11444455\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":68,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01422","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}