Home on Tian YE@HKUST,Guangzhou https://owen718.github.io/ Recent content in Home on Tian YE@HKUST,Guangzhou Hugo -- 0.146.0 en Sun, 12 Jul 2026 00:19:00 +0800 How a Coding Agent Reached the Global Top 10 on AgentTank in Under 24 Hours https://owen718.github.io/blogs/from-zero-to-global-top-10-agent-validation-pipeline/ Sun, 12 Jul 2026 00:19:00 +0800 https://owen718.github.io/blogs/from-zero-to-global-top-10-agent-validation-pipeline/ A practical account of building an evolvable validation pipeline that took an AgenTank coding agent from zero points to the global Top 10. 一个 Coding Agent 如何在 24 小时内冲进 AgentTank 全球 Top 10 https://owen718.github.io/blogs/from-zero-to-global-top-10-agent-validation-pipeline-zh/ Sun, 12 Jul 2026 00:19:00 +0800 https://owen718.github.io/blogs/from-zero-to-global-top-10-agent-validation-pipeline-zh/ 从零分到全球 Top 10：一套面向 Coding Agent 的可进化验证管线实践。 GenEvolve: Self-Evolving Image Generation Agents via Tool-Orchestrated Visual Experience Distillation https://owen718.github.io/pubs/41-2026-arxiv-genevolve/ Wed, 20 May 2026 00:00:00 +0000 https://owen718.github.io/pubs/41-2026-arxiv-genevolve/ SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformer https://owen718.github.io/pubs/39-2026-arxiv-sanawm/ Thu, 14 May 2026 00:00:00 +0000 https://owen718.github.io/pubs/39-2026-arxiv-sanawm/ An open-source world model for controllable minute-long 720p video generation on a single GPU. Building Real-Time Editing on FLUX2: Inference Acceleration and Distillation with Reinforcement Learning (Preview) https://owen718.github.io/blogs/realtime-editing-on-flux2/ Fri, 20 Mar 2026 23:08:56 +0800 https://owen718.github.io/blogs/realtime-editing-on-flux2/ A preview of a FLUX2 real-time editing stack covering inference acceleration, causal-editing distillation, and DMDR with an editing reward. LucidNFT: LR-Anchored Multi-Reward Preference Optimization for Flow-Based Real-World Super-Resolution https://owen718.github.io/pubs/40-2026-arxiv-lucidnft/ Fri, 06 Mar 2026 00:00:00 +0000 https://owen718.github.io/pubs/40-2026-arxiv-lucidnft/ UltraFlux: Data-Model Co-Design for High-quality Native 4K Text-to-Image Generation across Diverse Aspect Ratios https://owen718.github.io/pubs/36-2026-arxiv-ultraflux/ Sat, 21 Feb 2026 00:00:00 +0000 https://owen718.github.io/pubs/36-2026-arxiv-ultraflux/ Native 4K text-to-image generation across diverse aspect ratios, matching or surpassing Seedream 4.0. LucidFlux: Caption-Free Universal Image Restoration via a Large-Scale Diffusion Transformer https://owen718.github.io/pubs/27-2026-arxiv-lucidflux/ Fri, 26 Sep 2025 00:00:00 +0000 https://owen718.github.io/pubs/27-2026-arxiv-lucidflux/ A caption-free 14B restoration model delivering photorealistic results beyond leading commercial systems. GlassWizard: Harvesting Diffusion Priors for Glass Surface Detection https://owen718.github.io/pubs/05-2025-iccv-glasswizard/ Thu, 01 May 2025 00:00:00 +0000 https://owen718.github.io/pubs/05-2025-iccv-glasswizard/ MagicInfinite: Generating Infinite Talking Videos with Your Words and Voice https://owen718.github.io/pubs/38-2025-techreport-magicinfinite/ Tue, 01 Apr 2025 00:00:00 +0000 https://owen718.github.io/pubs/38-2025-techreport-magicinfinite/ Technical report for Hedra Inc.’s Character-3 model. AGLLDiff: Guiding Diffusion Models Towards Unsupervised Training-free Real-world Low-light Image Enhancement https://owen718.github.io/pubs/12-2025-aglldiff/ Sat, 01 Feb 2025 00:00:00 +0000 https://owen718.github.io/pubs/12-2025-aglldiff/ PromptHaze: Prompting Real-world Dehazing via Depth Anything Model https://owen718.github.io/pubs/13-2025-prompthaze/ Sat, 01 Feb 2025 00:00:00 +0000 https://owen718.github.io/pubs/13-2025-prompthaze/ Learning Diffusion Texture Priors for Image Restoration https://owen718.github.io/pubs/17-2024-cvpr-dtpm/ Sat, 01 Jun 2024 00:00:00 +0000 https://owen718.github.io/pubs/17-2024-cvpr-dtpm/ Meissonic: Revitalizing Masked Generative Transformers for Efficient High-Resolution Text-to-Image Synthesis https://owen718.github.io/pubs/11-2025-meissonic/ Mon, 01 Jan 2024 00:00:00 +0000 https://owen718.github.io/pubs/11-2025-meissonic/ The first SDXL-level high-resolution non-AR T2I model. Adverse Weather Removal with Codebook Priors https://owen718.github.io/pubs/20-2023-iccv-codebook/ Sun, 01 Oct 2023 00:00:00 +0000 https://owen718.github.io/pubs/20-2023-iccv-codebook/ Sparse Sampling Transformer with Uncertainty-Driven Ranking for Unified Removal of Raindrops and Rain Streaks https://owen718.github.io/pubs/19-2023-iccv-udr/ Sun, 01 Oct 2023 00:00:00 +0000 https://owen718.github.io/pubs/19-2023-iccv-udr/ Perceiving and Modeling Density for Image Dehazing https://owen718.github.io/pubs/32-2022-eccv-dest/ Sat, 01 Oct 2022 00:00:00 +0000 https://owen718.github.io/pubs/32-2022-eccv-dest/ Squint-to-Drag: Exploring Hands-free Eye-based Drag-and-Drop Interaction https://owen718.github.io/projects/02-squint/ Sat, 01 Feb 2020 00:00:00 +0000 https://owen718.github.io/projects/02-squint/ Dancing Lights: Supporting Dance Performances for Deaf and Hard of Hearing People through Lighting Effects https://owen718.github.io/projects/01-dance-lights/ Sat, 09 Mar 2019 00:00:00 +0000 https://owen718.github.io/projects/01-dance-lights/