1<!doctype html> 2<html lang="en" class="no-js"> 3 <head> 4 <meta charset="utf-8"> 5 6<!-- begin SEO --> 7 8 9 10 11 12<title>Ximing Xing - Homepage</title> 13 14 15 16 17 18 19 20<meta property="og:locale" content="en"> 21<meta property="og:site_name" content="Ximing Xing"> 22<meta property="og:title" content="Ximing Xing"> 23 24 25 <link rel="canonical" href="https://ximinng.github.io/"> 26 <meta property="og:url" content="https://ximinng.github.io/"> 27 28 29 30 <meta property="og:description" content="Ph.D. student at BUAA (Beihang University). Interested in Deep Generation and Vector Art."> 31 32 33 34 35 36 37 38 39 40<!-- end SEO --> 41 42 43<meta name="HandheldFriendly" content="True"> 44<meta name="MobileOptimized" content="320"> 45<meta name="viewport" content="width=device-width, initial-scale=1.0"> 46
47<script> 48 document.documentElement.className = document.documentElement.className.replace(/\bno-js\b/g, '') + ' js '; 49</script>
49 50 51<!-- For all browsers --> 52<link rel="stylesheet" href="assets/css/main.css"> 53 54<meta http-equiv="cleartype" content="on"> 55<head> 56 <base target="_blank"> 57</head> 58 <link rel="apple-touch-icon" sizes="180x180" href="images/apple-touch-icon.png"> 59<link rel="icon" type="image/png" sizes="32x32" href="images/favicon-32x32.png"> 60<link rel="icon" type="image/png" sizes="16x16" href="images/favicon-16x16.png"> 61<link rel="manifest" href="images/site.webmanifest"> 62 63<meta name="msapplication-TileColor" content="#000000"> 64<meta name="msapplication-TileImage" content="images/mstile-144x144.png?v=M44lzPylqQ"> 65<meta name="msapplication-config" content="images/browserconfig.xml?v=M44lzPylqQ"> 66<meta name="theme-color" content="#ffffff"> 67<link rel="stylesheet" href="assets/css/academicons.css"/> 68
69<script type="text/x-mathjax-config"> MathJax.Hub.Config({ TeX: { equationNumbers: { autoNumber: "all" } } }); </script>
vendor: 1 bytes, line 69
69
70<script type="text/x-mathjax-config"> 71 MathJax.Hub.Config({ 72 tex2jax: { 73 inlineMath: [ ['$','$'], ["\\(","\\)"] ], 74 processEscapes: true 75 } 76 }); 77</script>
vendor: 1 bytes, line 77
77
78<script src='https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.4/latest.js?config=TeX-MML-AM_CHTML' async></script>
78 79 80<!-- end custom head snippets --> 81 82 </head> 83 84 <body> 85 86 <!--[if lt IE 9]> 87<div class="notice--danger align-center" style="margin: 0;">You are using an <strong>outdated</strong> browser. Please <a href="http://browsehappy.com/">upgrade your browser</a> to improve your experience.</div> 88<![endif]--> 89 <div class="masthead"> 90 <div class="masthead__inner-wrap"> 91 <div class="masthead__menu"> 92 <nav id="site-nav" class="greedy-nav"> 93 <button><div class="navicon"></div></button> 94 <ul class="visible-links"> 95 <li class="masthead__menu-item masthead__menu-item--lg masthead__menu-home-item"><a href="#about-me">Homepage</a></li> 96 97 <li class="masthead__menu-item"><a href="/#about-me">About Me</a></li> 98 99 <li class="masthead__menu-item"><a href="/#-news">News</a></li> 100 101 <li class="masthead__menu-item"><a href="/#-publications">Publications</a></li> 102 103 <li class="masthead__menu-item"><a href="/#-projects">Projects</a></li> 104 105 <li class="masthead__menu-item"><a href="/#-honors-and-awards">Honors and Awards</a></li> 106 107 <li class="masthead__menu-item"><a href="/#-internships">Internships</a></li> 108 109 <li class="masthead__menu-item"><a href="/#-professional-activities">Professional Activities</a></li> 110 111 <li class="masthead__menu-item"><a href="/#-invited-talk">Invited Talk</a></li> 112 113 </ul> 114 <ul class="hidden-links hidden"></ul> 115 </nav> 116 </div> 117 </div> 118</div> 119 120 <div id="main" role="main"> 121 122 <div class="sidebar sticky"> 123 124 125<div itemscope itemtype="http://schema.org/Person" class="profile_box"> 126 127 <div class="author__avatar"> 128 <img src="images/profile.png" class="author__avatar" alt="Ximing Xing"> 129 </div> 130 131 <div class="author__content"> 132 <h3 class="author__name">Ximing Xing</h3> 133 <p class="author__bio">BeiHang University</p> 134 </div> 135 136 <div class="author__urls-wrapper"> 137 <!-- <button class="btn btn--inverse">More Info & Contact</button> --> 138 <ul class="author__urls social-icons"> 139 140 <li><div style="white-space: normal; margin-bottom: 1em;">Ph.D. student at BUAA (Beihang University). Interested in Deep Generation and Vector Art.</div></li> 141 142 143 <li><i class="fa fa-fw fa-map-marker" aria-hidden="true"></i> Beijing, China</li> 144 145 146 147 148 <li><a href="mailto:[email protected]"><i class="fas fa-fw fa-envelope" aria-hidden="true"></i> Email</a></li> 149 150 151 152 153 154 155 156 <li><a href="https://huggingface.co/xingxm"><img src="https://huggingface.co/front/assets/huggingface_logo-noborder.svg" style="width:1.1em;height:1.1em;vertical-align:-0.2em;margin-right:4px;" alt="HuggingFace"> HuggingFace</a></li> 157 158 159 160 161 162 163 164 165 <li><a href="https://github.com/ximinng"><i class="fab fa-fw fa-github" aria-hidden="true"></i> Github</a></li> 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 <li><a href="https://scholar.google.com/citations?user=tFpaF7AAAAAJ"><i class="fas fa-fw fa-graduation-cap"></i> Google Scholar</a></li> 181 182 183 184 185 186 </ul> 187 <div class="author__urls_sm"> 188 189 190 <a href="mailto:[email protected]"><i class="fas fa-fw fa-envelope" aria-hidden="true"></i></a> 191 192 193 194 195 196 197 198 <a href="https://huggingface.co/xingxm"><img src="https://huggingface.co/front/assets/huggingface_logo-noborder.svg" style="width:1.1em;height:1.1em;vertical-align:-0.2em;" alt="HuggingFace"></a> 199 200 201 202 203 204 205 206 207 <a href="https://github.com/ximinng"><i class="fab fa-fw fa-github" aria-hidden="true"></i></a> 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 <a href="https://scholar.google.com/citations?user=tFpaF7AAAAAJ"><i class="fas fa-fw fa-graduation-cap"></i></a> 223 224 225 226 227 228 </div> 229 </div> 230</div> 231 232 233 </div> 234 235 236 <article class="page" itemscope itemtype="http://schema.org/CreativeWork"> 237 <meta itemprop="headline" content="Ximing Xing, XiMing Xing, ximingxing, è¡ä¹ é, ximing xing, ximing, ximinng"> 238 <div class="page__inner-wrap"> 239 <section class="page__content" itemprop="text"> 240 241<p><span class="anchor" id="about-me"></span></p> 242 243<p>Ximing Xing is a Ph.D. student (2022-Present) in Software Engineering at Beihang University, advised by Professor <a href="https://yuqian1023.github.io/">Qian Yu</a>. He is currently a research intern at Tencent Hunyuan, working on multimodal vector graphics large language models.</p> 244 245<p>He is dedicated to advancing the frontiers of AI-driven content creation, with a particular focus on the generation and 246understanding of vector graphics. His research spans <strong>deep generative models, vector art synthesis, text-to-SVG generation, neural rendering, SVG diffusion models, and multimodal SVG LLMs</strong>.</p> 247 248<p>His work has been published in top-tier venues including <strong>CVPRâ24/25/26, T-PAMIâ25, and NeurIPSâ23</strong>, with his <strong>T-PAMI paper being the first SVG generation work ever published</strong> in this prestigious journal. 249Some of his representative works include:</p> 250 251<ul> 252 <li><strong>LLMs for Vector Graphics</strong>: <a href="https://ximinng.github.io/LLM4SVGProject/">LLM4SVG</a> <a href="https://github.com/ximinng/LLM4SVG"><img src="https://img.shields.io/github/stars/ximinng/LLM4SVG?style=social&label=Code" alt="" /></a>, <a href="https://arxiv.org/abs/2505.24499">Reason-SVG</a></li> 253 <li><strong>Text-to-SVG Synthesis</strong>: <a href="https://ximinng.github.io/SVGDreamerV2Project/">SVGDreamer++</a> <a href="https://github.com/ximinng/SVGDreamerV2"><img src="https://img.shields.io/github/stars/ximinng/SVGDreamerV2?style=social&label=Code" alt="" /></a>, <a href="https://ximinng.github.io/SVGDreamer-project/">SVGDreamer</a> <a href="https://github.com/ximinng/SVGDreamer"><img src="https://img.shields.io/github/stars/ximinng/SVGDreamer?style=social&label=Code" alt="" /></a></li> 254 <li><strong>Vector Diffusion Models</strong>: <a href="https://ximinng.github.io/SVGFusionProject/">SVGFusion</a> <a href="https://github.com/ximinng/SVGFusion"><img src="https://img.shields.io/github/stars/ximinng/SVGFusion?style=social&label=Code" alt="" /></a></li> 255 <li><strong>Sketch Synthesis</strong>: <a href="https://ximinng.github.io/DiffSketcher-project/">DiffSketcher</a> <a href="https://github.com/ximinng/DiffSketcher"><img src="https://img.shields.io/github/stars/ximinng/DiffSketcher?style=social&label=Code" alt="" /></a></li> 256 <li><strong>Differentiable SVG Rendering Library</strong>: <a href="https://ximinng.github.io/PyTorch-SVGRender-project/">PyTorch-SVGRender</a> <a href="https://github.com/ximinng/PyTorch-SVGRender"><img src="https://img.shields.io/github/stars/ximinng/PyTorch-SVGRender?style=social&label=Code" alt="" /></a></li> 257</ul> 258 259<p>Beyond academic publications, he is committed to open science and reproducible research â actively contributing <strong>large-scale datasets</strong>, <strong>production-ready code libraries</strong>, and <strong>interactive demos</strong>
259 on HuggingFace (<a href="https://huggingface.co/xingxm">@xingxm</a>, <a href="https://huggingface.co/SVGRender">SVGRender Space</a>), making cutting-edge vector graphics research accessible to the broader community. 260<a href="https://github.com/ximinng/"><img src="https://img.shields.io/badge/dynamic/json?logo=github&label=GitHub%20Stars&style=for-the-badge&query=%24.stars&url=https://api.github-star-counter.workers.dev/user/ximinng" alt="github" /></a></p> 261 262<h1 id="-news">ð¥ News</h1> 263 264<ul> 265 <li><em>2026.03</em>:  ðð Two papers <a href="https://hunyuan-promptenhancer.github.io/">PromptEnhancer</a> <a href="https://github.com/Hunyuan-PromptEnhancer/PromptEnhancer"><img src="https://img.shields.io/github/stars/Hunyuan-PromptEnhancer/PromptEnhancer?style=social&label=Code+Stars" alt="" /></a> and <a href="https://arxiv.org/abs/2505.24499">Reason-SVG</a> have been accepted by <strong>CVPR 2026!</strong></li> 266 <li><em>2025.02</em>:  ðð Our paper <a href="https://ximinng.github.io/LLM4SVGProject/">LLM4SVG</a> has been accepted by <strong>CVPR 2025!</strong></li> 267 <li><em>2025.02</em>:  ðð Our paper <a href="https://ximinng.github.io/SVGDreamerV2Project/">SVGDreamer++</a> has been accepted by <strong>T-PAMI!</strong> This is the <strong>first paper on SVG generation</strong> ever published in <strong>T-PAMI</strong>.</li> 268 <li><em>2024.02</em>:  ðð Our paper <a href="https://ximinng.github.io/SVGDreamer-project/">SVGDreamer</a> has been accepted by <strong>CVPR 2024!</strong></li> 269 <li><em>2023.12</em>:  ðð We released <a href="https://github.com/ximinng/PyTorch-SVGRender">PyTorch-SVGRender</a>, a state-of-the-art library for differentiable SVG rendering in PyTorch.</li> 270</ul> 271 272<h1 id="-publications">ð Publications</h1> 273 274<div class="pub-filter-bar"> 275 <button class="pub-filter-btn active" onclick="filterPubs('selected', this)">â Selected Papers <span class="btn-count">4</span></button> 276 <button class="pub-filter-btn" onclick="filterPubs('vgs', this)">ð¨ Vector Graphics Synthesis <span class="btn-count">7</span></button> 277 <button class="pub-filter-btn" onclick="filterPubs('cad', this)">ð© CAD Generation <span class="btn-count">1</span></button> 278 <button class="pub-filter-btn" onclick="filterPubs('t2i', this)">ð¼ï¸ Controllable T2I Generation <span class="btn-count">2</span></button> 279 <button class="pub-filter-btn" onclick="filterPubs('rml', this)">ð¡ï¸ Robust Machine Learning <span class="btn-count">1</span></button> 280</div> 281 282<!-- ===== SELECTED PAPERS ===== --> 283<div id="pub-selected" class="pub-topic-group visible"> 284 285<div class="paper-box"> 286<div class="paper-box-image"><div><div class="badge">T-PAMI 2025</div><img src="images/svgdreamerv2_cover.png" loading="lazy" alt="SVGDreamer++" /></div></div> 287<div class="paper-box-text"> 288 <p><a href="https://arxiv.org/abs/2411.17832">SVGDreamer++: Advancing Editability and Diversity in Text-Guided SVG Generation</a></p> 289 290 <p><strong>Ximing Xing</strong>, Qian Yu, Chuang Wang, Haitao Zhou, Jing Zhang, Dong Xu</p> 291 292 <p><a href="https://ximinng.github.io/SVGDreamerV2Project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-SVGDreamer++-orange.svg" alt="project" /></a> <a href="https://github.com/ximinng/SVGDreamerV2"><img src="https://img.shields.io/github/stars/ximinng/SVGDreamerV2?style=social&label=Code+Stars" alt="" /></a></p> 293 294 <p><b><u>TL;DR:</u></b> SVGDreamer++ is an advanced text-to-SVG generator with two core innovations: 295<strong>Hierarchical Image Vectorization (HIVE)</strong> - enables semantic object-level and component-level image vectorization, 296and <strong>Adaptive Vector Primitive Control</strong> â dynamically assigns the optimal number of vector primitives, capturing fine-grained details without wasting computation.</p> 297 298 <p class="conf-color">IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI 2025)</p> 299 300 <p><a href="https://ximinng.github.io/SVGDreamerV2Project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/SVGDreamerV2"><strong>Code</strong></a> | <a href="https://huggingface.co/blog/xingxm/svgdreamer"><strong>Blog</strong></a></p> 301 </div> 302</div> 303 304<div class="paper-box"> 305<div class="paper-box-image"><div><div class="badge">CVPR 2026</div><img src="images/reasonsvg_cover.png" loading="lazy" alt="ReasonSVG" /></div></div> 306<div class="paper-box-text"> 307 <p><a href="https://arxiv.org/abs/2505.24499">Reason-SVG: Enhancing Structured Reasoning for Scalable Vector Graphics Generation with Reinforcement Learning</a></p> 308 309 <p><strong>Ximing Xing</strong>, Ziteng Xue, Yandong Guan, Jing Zhang, Dong Xu, Qian Yu</p> 310 311 <p><b><u>TL;DR:</u></b> Reason-SVG introduces the first framework to enhance SVG generation in LLMs through a <strong>âDrawing-with-Thoughtâ (DwT)</strong> paradigmâcombining explicit design reasoning with codeâtrained via supervised fine-tuning and <strong>HyperReward-driven</strong> reinforcement learning.</p> 312 313 </div> 314</div> 315 316<div class="paper-box"> 317<div class="paper-box-image"><div><div class="badge">CVPR 2024</div><img src="images/svgdreamer11_cover.png" loading="lazy" alt="SVGDreamer" /></div></div> 318<div class="paper-box-text"> 319 <p><a href="https://arxiv.org/abs/2312.16476">SVGDreamer: Text Guided SVG Generation with Diffusion Model</a></p> 320 321 <p><strong>Ximing Xing</strong>, Chuang Wang, Haitao Zhou, Jing Zhang, Dong Xu, Qian Yu</p> 322 323 <p><a href="https://ximinng.github.io/SVGDreamer-project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-SVGDreamer-orange.svg" alt="project" /></a> <a href="https://github.com/ximinng/SVGDreamer"><img src="https://img.shields.io/github/stars/ximinng/SVGDreamer?style=social&label=Code+Stars" alt="" /></a></p> 324 325 <p><b><u>TL;DR:</u></b> SVGDreamer introduces <strong>Semantic-driven Image VEctorization (SIVE)</strong> and <strong>
325Vector Particle-based Score Distillation (VPSD)</strong> to generate editable, high-quality SVGs with better shape control and diversity.</p> 326 327 <p class="conf-color">IEEE Conference on Computer Vision and Pattern Recognition (CVPR'24)</p> 328 329 <p><a href="https://ximinng.github.io/SVGDreamer-project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/SVGDreamer"><strong>Code</strong></a> | <a href="https://huggingface.co/blog/xingxm/svgdreamer"><strong>Blog</strong></a></p> 330 </div> 331</div> 332 333<div class="paper-box"> 334<div class="paper-box-image"><div><div class="badge">NIPS 2023</div><img src="images/diffsketcher_cover.png" loading="lazy" alt="DiffSketcher" /></div></div> 335<div class="paper-box-text"> 336 <p><a href="https://arxiv.org/abs/2306.14685">DiffSketcher: Text Guided Vector Sketch Synthesis through Latent Diffusion Models</a></p> 337 338 <p><strong>Ximing Xing</strong>, Chuang Wang, Haitao Zhou, Jing Zhang, Qian Yu, Dong Xu</p> 339 340 <p><a href="https://ximinng.github.io/DiffSketcher-project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-DiffSketcher-orange.svg" alt="project" /></a> <a href="https://github.com/ximinng/DiffSketcher"><img src="https://img.shields.io/github/stars/ximinng/DiffSketcher?style=social&label=Code+Stars" alt="" /></a></p> 341 342 <p><b><u>TL;DR:</u></b> DiffSketcher pioneered the use of diffusion models for <strong>text-to-vector sketch</strong> generation.</p> 343 344 <p class="conf-color">Advances in Neural Information Processing Systems (NeurIPS'23)</p> 345 346 <p><a href="https://ximinng.github.io/DiffSketcher-project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/DiffSketcher"><strong>Code</strong></a></p> 347 </div> 348</div> 349 350</div> 351 352<!-- ===== TOPIC: Vector Graphics Synthesis ===== --> 353<div id="pub-vgs" class="pub-topic-group"> 354 355<div class="pub-topic-header"> 356 <span class="pub-topic-title">ð¨ Vector Graphics Generation</span> 357 <span class="pub-topic-count-badge">7 papers</span> 358</div> 359 360<div class="paper-box"> 361<div class="paper-box-image"><div><div class="badge">CVPR 2026</div><img src="images/reasonsvg_cover.png" loading="lazy" alt="ReasonSVG" /></div></div> 362<div class="paper-box-text"> 363 <p><a href="https://arxiv.org/abs/2505.24499">Reason-SVG: Enhancing Structured Reasoning for Scalable Vector Graphics Generation with Reinforcement Learning</a></p> 364 365 <p><strong>Ximing Xing</strong>, Ziteng Xue, Yandong Guan, Jing Zhang, Dong Xu, Qian Yu</p> 366 367 <p><b><u>TL;DR:</u></b> Reason-SVG introduces the first framework to enhance SVG generation in LLMs through a <strong>âDrawing-with-Thoughtâ (DwT)</strong> paradigmâcombining explicit design reasoning with codeâtrained via supervised fine-tuning and <strong>HyperReward-driven</strong> reinforcement learning.</p> 368 369 </div> 370</div> 371 372<div class="paper-box"> 373<div class="paper-box-image"><div><div class="badge">CVPR 2025</div><img src="images/llm4svg_cover.png" loading="lazy" alt="LLM4SVG" /></div></div> 374<div class="paper-box-text"> 375 <p><a href="https://arxiv.org/abs/2412.11102">Empowering LLMs to Understand and Generate Complex Vector Graphics</a></p> 376 377 <p><strong>Ximing Xing</strong>, Juncheng Hu, Guotao Liang, Jing Zhang, Dong Xu, Qian Yu</p> 378 379 <p><a href="https://ximinng.github.io/LLM4SVGProject/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-LLM4SVG-orange.svg" alt="project" /></a> <a href="https://huggingface.co/datasets/xingxm/SVGX-SFT-1M"><img src="https://img.shields.io/badge/Dataset-SVGX_SFT_1M-ffcc00?logo=huggingface" alt="dataset" /></a> <a href="https://github.com/ximinng/LLM4SVG"><img src="https://img.shields.io/github/stars/ximinng/LLM4SVG?style=social&label=Code+Stars" alt="" /></a></p> 380 381 <p><b><u>TL;DR:</u></b> LLM4SVG introduces learnable <strong>SVG Semantic Tokens</strong> and a large <strong>SVGX-SFT dataset</strong>, enabling LLMs to understand and generate complex vector graphics.</p> 382 383 <p class="conf-color">IEEE Conference on Computer Vision and Pattern Recognition (CVPR'25)</p> 384 385 <p><a href="https://ximinng.github.io/LLM4SVGProject/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/LLM4SVG"><strong>Code</strong></a> | <a href="https://huggingface.co/datasets/xingxm/SVGX-SFT-1M"><strong>SVGX-SFT-1M Dataset</strong></a></p> 386 </div> 387</div> 388 389<div class="paper-box"> 390<div class="paper-box-image"><div><div class="badge">T-PAMI 2025</div><img src="images/svgdreamerv2_cover.png" loading="lazy" alt="SVGDreamer++" /></div></div> 391<div class="paper-box-text"> 392 <p><a href="https://arxiv.org/abs/2411.17832">SVGDreamer++: Advancing Editability and Diversity in Text-Guided SVG Generation</a></p> 393 394 <p><strong>Ximing Xing</strong>, Qian Yu, Chuang Wang, Haitao Zhou, Jing Zhang, Dong Xu</p> 395 396 <p><a href="https://ximinng.github.io/SVGDreamerV2Project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-SVGDreamer++-orange.svg" alt="project" /></a> <a href="https://github.com/ximinng/SVGDreamerV2"><img src="https://img.shields.io/github/stars/ximinng/SVGDreamerV2?style=social&label=Code+Stars" alt="" /></a></p> 397 398 <p><b><u>TL;DR:</u></b> SVGDreamer++ is an advanced text-to-SVG generator with two core innovations: 399<strong>Hierarchical Image Vectorization (HIVE)</strong> - enables semantic object-level and component-level image vectorization, 400and <strong>Adaptive Vector Primitive Control</strong> â dynamically assigns the optimal number of vector primitives, capturing fine-grained details without wasting computation.</p> 401 402 <p class="conf-color">IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI 2025)</p> 403 404 <p><a href="https://ximinng.github.io/SVGDreamerV2Project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/SVGDreamerV2"><strong>Code</strong></a> | <a href="https://huggingface.co/blog/xingxm/svgdreamer"><strong>Blog</strong></a></p> 405 </div> 406</div> 407 408<div class="paper-box"> 409<div class="paper-box-image"><div><div class="badge">arXiv 2024</div><img src="images/svgfusion_cover.png" loading="lazy" alt="SVGFusion" /></div></div> 410<div class="paper-box-text"> 411 <p><a href="https://arxiv.org/abs/2412.10437">SVGFusion: Scalable Text-to-SVG Generation via Vector Space Diffusion</a></p> 412 413 <p><strong>Ximing Xing</strong>, Juncheng Hu, Jing Zhang, Dong Xu, Qian Yu</p> 414 415 <p><a href="https://ximinng.github.io/SVGFusionProject/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-SVGFusion-orange.svg" alt="project" /></a> <a href="https://huggingface.co/datasets/xingxm/SVGX-Core-250k"><img src="https://img.shields.io/badge/Dataset-SVGX_Core_250k-ffcc00?logo=huggingface" alt="dataset" /></a> <a href="https://github.com/ximinng/SVGFusion"><img src="https://img.shields.io/github/stars/ximinng/SVGFusion?style=social&label=Code+Stars" alt="" /></a></p> 416 417 <p><b><u>TL;DR:</u></b> SVGFusion improves text-to-SVG generation by using a <strong>VP-VAE to learn a vector representation of SVG elements</strong>, and a <strong>VS-DiT</strong> to generate SVGs from text prompts by performing diffusion within that <strong>learned vector space</strong>.</p> 418 419 <p><a href="https://ximinng.github.io/SVGFusionProject/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/SVGFusion"><strong>Code</strong></a> | <a href="https://huggingface.co/datasets/xingxm/SVGX-Core-250k"><strong>SVGX-Core-250k Dataset</strong></a></p> 420 </div> 421</div> 422 423<div class="paper-box"> 424<div class="paper-box-image"><div><div class="badge">CVPR 2024</div><img src="images/svgdreamer11_cover.png" loading="lazy" alt="SVGDreamer" /></div></div> 425<div class="paper-box-text"> 426 <p><a href="https://arxiv.org/abs/2312.16476">SVGDreamer: Text Guided SVG Generation with Diffusion Model</a></p> 427 428 <p><strong>Ximing Xing</strong>, Chuang Wang, Haitao Zhou, Jing Zhang, Dong Xu, Qian Yu</p> 429 430 <p><a href="https://ximinng.github.io/SVGDreamer-project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-SVGDreamer-orange.svg" alt="project" /></a> <a href="https://github.com/ximinng/SVGDreamer"><img src="https://img.shields.io/github/stars/ximinng/SVGDreamer?style=social&label=Code+Stars" alt="" /></a></p> 431 432 <p><b><u>TL;DR:</u></b> SVGDreamer introduces <strong>Semantic-driven Image VEctorization (SIVE)</strong> and <strong>
432Vector Particle-based Score Distillation (VPSD)</strong> to generate editable, high-quality SVGs with better shape control and diversity.</p> 433 434 <p class="conf-color">IEEE Conference on Computer Vision and Pattern Recognition (CVPR'24)</p> 435 436 <p><a href="https://ximinng.github.io/SVGDreamer-project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/SVGDreamer"><strong>Code</strong></a> | <a href="https://huggingface.co/blog/xingxm/svgdreamer"><strong>Blog</strong></a></p> 437 </div> 438</div> 439 440<div class="paper-box"> 441<div class="paper-box-image"><div><div class="badge">NIPS 2023</div><img src="images/diffsketcher_cover.png" loading="lazy" alt="DiffSketcher" /></div></div> 442<div class="paper-box-text"> 443 <p><a href="https://arxiv.org/abs/2306.14685">DiffSketcher: Text Guided Vector Sketch Synthesis through Latent Diffusion Models</a></p> 444 445 <p><strong>Ximing Xing</strong>, Chuang Wang, Haitao Zhou, Jing Zhang, Qian Yu, Dong Xu</p> 446 447 <p><a href="https://ximinng.github.io/DiffSketcher-project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-DiffSketcher-orange.svg" alt="project" /></a> <a href="https://github.com/ximinng/DiffSketcher"><img src="https://img.shields.io/github/stars/ximinng/DiffSketcher?style=social&label=Code+Stars" alt="" /></a></p> 448 449 <p><b><u>TL;DR:</u></b> DiffSketcher pioneered the use of diffusion models for <strong>text-to-vector sketch</strong> generation.</p> 450 451 <p class="conf-color">Advances in Neural Information Processing Systems (NeurIPS'23)</p> 452 453 <p><a href="https://ximinng.github.io/DiffSketcher-project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/DiffSketcher"><strong>Code</strong></a></p> 454 </div> 455</div> 456 457<div class="paper-box"> 458<div class="paper-box-image"><div><div class="badge">ICME 2025</div><img src="images/vectorpainter-cover.png" loading="lazy" alt="VectorPainter" /></div></div> 459<div class="paper-box-text"> 460 <p><a href="https://arxiv.org/abs/2405.02962">VectorPainter: Advanced Stylized Vector Graphics Synthesis Using Stroke-Style Priors</a></p> 461 462 <p>Juncheng Hu, <strong>Ximing Xing</strong>, Jing Zhang, Qian Yu</p> 463 464 <p><a href="https://hjc-owo.github.io/VectorPainterProject/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-VectorPainter-orange.svg" alt="project" /></a></p> 465 466 <p><b><u>TL;DR:</u></b> VectorPainter synthesizes text-guided vector graphics by <strong>imitating stylized strokes</strong>.</p> 467 468 <p class="conf-color">IEEE International Conference on Multimedia & Expo (ICME'25)</p> 469 470 <p><a href="https://hjc-owo.github.io/VectorPainterProject/"><strong>Project</strong></a> | <a href="https://github.com/hjc-owo/VectorPainter"><strong>Code</strong></a></p> 471 </div> 472</div> 473 474</div> 475 476<!-- ===== TOPIC: CAD Generation ===== --> 477<div id="pub-cad" class="pub-topic-group"> 478 479<div class="pub-topic-header"> 480 <span class="pub-topic-title">ð© Parametric generation</span> 481 <span class="pub-topic-count-badge">1 paper</span> 482</div> 483 484<div class="paper-box"> 485<div class="paper-box-image"><div><div class="badge">NeurIPS 2025</div><img src="images/cadcoder_cover.png" loading="lazy" alt="CAD-Coder" /></div></div> 486<div class="paper-box-text"> 487 488 <p><a href="https://arxiv.org/abs/2505.19713">CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward</a></p> 489 490 <p>Yandong Guan, Xilin Wang, <strong>Ximing Xing</strong>, Jing Zhang, Dong Xu, Qian Yu</p> 491 492 <p><b><u>TL;DR:</u></b> CAD-Coder enables LLMs to generate complex, valid 3D CAD models from text by outputting CadQuery (Python) scripts, using a novel structured chain-of-thought approach.</p> 493 494 <p><a href="https://github.com/gudo7208/CAD-Coder"><strong>Code</strong></a> | <a href="https://huggingface.co/gudo7208/CAD-Coder"><strong>Model</strong></a> | <a href="https://huggingface.co/datasets/gudo7208/CAD-Coder"><strong>Dataset</strong></a></p> 495 </div> 496</div> 497 498</div> 499 500<!-- ===== TOPIC: Controllable Text-to-Image Generation ===== --> 501<div id="pub-t2i" class="pub-topic-group"> 502 503<div class="pub-topic-header"> 504 <span class="pub-topic-title">ð¼ï¸ Controllable Text-to-Image Generation</span> 505 <span class="pub-topic-count-badge">2 papers</span> 506</div> 507 508<div class="paper-box"> 509<div class="paper-box-image"><div><div class="badge">CVPR 2026</div><img src="images/promptenhancer_cover.png" loading="lazy" alt="PromptEnhancer" /></div></div> 510<div class="paper-box-text"> 511 512 <p><a href="https://arxiv.org/abs/2509.04545">PromptEnhancer: Taming Your Rewriter for Text-to-Image Generation via Fine-Grained Reward</a></p> 513 514 <p>Linqing Wang, <strong>Ximing Xing</strong> et al.</p> 515 516 <p><a href="https://hunyuan-promptenhancer.github.io/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-PromptEnhancer-orange.svg" alt="project" /></a> <a href="https://huggingface.co/datasets/PromptEnhancer/T2I-Keypoints-Eval"><img src="https://img.shields.io/badge/Benchmark-T2I_Keypoints-ffcc00?logo=huggingface" alt="benchmark" /></a> <a href="https://github.com/Hunyuan-PromptEnhancer/PromptEnhancer"><img src="https://img.shields.io/github/stars/Hunyuan-PromptEnhancer/PromptEnhancer?style=social&label=Code+Stars" alt="" /></a></p> 517 518 <p><b><u>TL;DR:</u></b> PromptEnhancer is a user prompt rewriting tool for enhancing text-to-image generation.</p> 519 520 <p><a href="https://hunyuan-promptenhancer.github.io/"><strong>Project</strong></a> | <a href="https://github.com/Hunyuan-PromptEnhancer/PromptEnhancer"><strong>Code</strong></a> | <a href="https://huggingface.co/tencent/HunyuanImage-2.1/tree/main/reprompt"><strong>
520Model</strong></a></p> 521 </div> 522</div> 523 524<div class="paper-box"> 525<div class="paper-box-image"><div><div class="badge">arXiv</div><img src="images/Inv-by-Inv-teaser.png" loading="lazy" alt="Inversion-By-Inversion" /></div></div> 526<div class="paper-box-text"> 527 528 <p><a href="https://arxiv.org/abs/2308.07665">Inversion-by-Inversion: Exemplar-based Sketch-to-Photo Synthesis via Stochastic Differential Equations without Training</a></p> 529 530 <p><strong>Ximing Xing</strong>, Chuang Wang, Haitao Zhou, Zhihao Hu, Chongxuan Li, Dong Xu, Qian Yu</p> 531 532 <p><a href="https://ximinng.github.io/inversion-by-inversion-project/"><img src="https://img.shields.io/badge/%F0%9F%8F%A0%20Homepage-InversionbyInversion-orange.svg" alt="project" /></a></p> 533 534 <p><b><u>TL;DR:</u></b> Free training for sketch control image synthesis via Stochastic Differential Equations (SDEs).</p> 535 536 <p><a href="https://ximinng.github.io/inversion-by-inversion-project/"><strong>Project</strong></a> | <a href="https://github.com/ximinng/inversion-by-inversion"><strong>Code</strong></a></p> 537 </div> 538</div> 539 540</div> 541 542<!-- ===== TOPIC: Robust Machine Learning ===== --> 543<div id="pub-rml" class="pub-topic-group"> 544 545<div class="pub-topic-header"> 546 <span class="pub-topic-title">ð¡ï¸ Robust Machine Learning</span> 547 <span class="pub-topic-count-badge">1 paper</span> 548</div> 549 550<div class="paper-box"> 551<div class="paper-box-image"><div><div class="badge">CVPR 2021</div><img src="images/dualgraph_cover.png" loading="lazy" alt="DualGraph" /></div></div> 552<div class="paper-box-text"> 553 554 <p><a href="https://openaccess.thecvf.com/content/CVPR2021/papers/Zhang_DualGraph_A_Graph-Based_Method_for_Reasoning_About_Label_Noise_CVPR_2021_paper.pdf">A Graph-Based Method for Reasoning About Label Noise</a></p> 555 556 <p>HaiYang Zhang, <strong>Ximing Xing</strong>, Liang Liu</p> 557 558 <p><b><u>TL;DR:</u></b> DualGraph, the first method for label noise processing based on graph neural networks.</p> 559 560 <p class="conf-color">IEEE Conference on Computer Vision and Pattern Recognition (CVPR'21)</p> 561 </div> 562</div> 563 564</div> 565 566<!-- [//]: # (# ð Educations) --> 567 568<!-- [//]: # (- *2022.09 - (now)*, Phd student, Beihang University, Beijing. ) --> 569<!-- - *2015.09 - 2019.06*, Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus ornare aliquet ipsum, ac tempus justo dapibus sit amet. --> 570 571<!-- # ð¬ Invited Talks 572- *2021.06*, Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus ornare aliquet ipsum, ac tempus justo dapibus sit amet. 573- *2021.03*, Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus ornare aliquet ipsum, ac tempus justo dapibus sit amet. \| [\[video\]](https://github.com/) --> 574 575<h1 id="-projects">ð Projects</h1> 576 577<!-- project 1 --> 578 579<div class="paper-box"> 580<div class="paper-box-image"><div><div class="project-badge">open source</div><img src="images/PyTorch-SVGRender-cover.png" alt="PyTorch-SVGRender" width="100%" /></div></div> 581<div class="paper-box-text"> 582 <p><a href="https://ximinng.github.io/PyTorch-SVGRender-project/">Pytorch-SVGRender: A Differentiable Rendering Library for SVG Creation</a></p> 583 584 <p>ð¥ Project Founder and Main Contributor: <strong>Ximing Xing</strong></p> 585 586 <p><b><u>TL;DR:</u></b> SVG Differentiable Rendering: Generating vector graphics using neural networks. Support: Text-to-SVG, Image-to-SVG and SVG Editing.</p> 587 588 <p><a href="https://ximinng.github.io/PyTorch-SVGRender-project/"><img src="https://img.shields.io/badge/Website-Gitpage-yellow" alt="website" /></a> 589<a href="https://pytorch-svgrender.readthedocs.io/en/latest/index.html"><img src="https://img.shields.io/badge/DOCS-Readthedocs-purple?logo=readthedocs" alt="docs" /></a> 590<a href="https://huggingface.co/SVGRender"><img src="https://img.shields.io/badge/SPACE-HuggingFace-ffcc00?logo=huggingface" alt="space" /></a> <a href="https://github.com/ximinng/PyTorch-SVGRender"><img src="https://img.shields.io/github/stars/ximinng/PyTorch-SVGRender?style=social&label=Code+Stars" alt="" /></a></p> 591 592 <p>ð <a href="https://ximinng.github.io/PyTorch-SVGRender-project/"><strong>Project</strong></a> | ð <a href="https://github.com/ximinng/PyTorch-SVGRender"><strong>Code</strong></a> | ð¤ <a href="https://huggingface.co/SVGRender"><strong>HuggingFace</strong></a> | ð <a href="https://pytorch-svgrender.readthedocs.io/en/latest/index.html"><strong>Docs</strong></a></p> 593 </div> 594</div> 595
596<h1 id="-honors--awards">ð Honors & Awards</h1> 597 598<ul> 599 <li><em>2025.04</em> Academic Excellence Foundation of BUAA for PhD Students.</li> 600 <li><em>2024.12</em> National Scholarship for Doctoral Students.</li> 601 <li><em>2021.12</em> National Scholarship.</li> 602</ul> 603 604<h1 id="-internships">ð» Internships</h1> 605 606<div style="border-left: 3px solid #0066cc; padding-left: 16px; margin-bottom: 28px;"> 607<p style="margin: 0 0 4px 0;"> 608 <strong>2025.06 â Present</strong> · 609 <a href="https://hunyuan.tencent.com/">Tencent Hunyuan</a> · 610 Research Intern, Qingyun Program · Beijing, China 611</p> 612<p style="margin: 6px 0 8px 0; font-weight: 600;">Multimodal Vector Graphics Large Language Models</p> 613<ul style="margin: 0;"> 614 <li>Developing novel approaches for multimodal vector graphics understanding and generation with LLMs.</li> 615 <li>Advancing the integration of large language models with vector graphics synthesis pipelines.</li> 616 <li>Leading research on SVG-native LLM architectures and reinforcement learning-based training strategies.</li> 617</ul> 618</div> 619 620<div style="border-left: 3px solid #aaa; padding-left: 16px; margin-bottom: 12px;"> 621<p style="margin: 0 0 4px 0;"> 622 <strong>2021.06 â 2021.11</strong> · 623 <a href="https://www.antgroup.com/en">Ant Group</a> · 624 Machine Learning Algorithm Intern · Hangzhou, China 625</p> 626<p style="margin: 6px 0 8px 0; font-weight: 600;">Multi-Turn Task-Oriented Dialogue System</p> 627<ul style="margin: 0;"> 628 <li>Designed behavior cloning strategies to enhance model cold-start training efficiency.</li> 629 <li>Built an agenda-based user simulator for reinforcement learning environments.</li> 630 <li>Engineered turn-level and session-level reward functions for PPO-based policy optimization.</li> 631</ul> 632</div> 633 634<h1 id="-professional-activities">ð Professional Activities</h1> 635 636<h4 id="conference-reviewer">Conference Reviewer</h4> 637 638<ul> 639 <li> 640 <p><strong>2025</strong>: AAAI, CVPR, SIGGRAPH, SIGGRAPH Asia, NeurIPS</p> 641 </li> 642 <li> 643 <p><strong>2024</strong>: CVPR, ECCV, NeurIPS, ACM MM</p> 644 </li> 645</ul> 646 647<h4 id="journal-reviewer">Journal Reviewer</h4> 648 649<ul> 650 <li><strong>Computer Vision & Graphics</strong>: International Journal of Computer Vision (IJCV), IEEE Transactions on Visualization and Computer Graphics (T-VCG)</li> 651</ul> 652 653<h4 id="academic-service">Academic Service</h4> 654 655<ul> 656 <li>Regular reviewer for top-tier conferences and journals in AI, Computer Vision, and Computer Graphics</li> 657</ul> 658 659 </section> 660 </div> 661 </article> 662 </div> 663 664
664<script src="assets/js/main.min.js"></script>
vendor: 1 bytes, line 664
664
665<script src="assets/js/pub-filter.js"></script>
665 666 667 668 669<!-- Global site tag (gtag.js) - Google Analytics -->
670<script async src="https://www.googletagmanager.com/gtag/js?id="></script>
vendor: 1 bytes, line 670
670
671<script> 672 window.dataLayer = window.dataLayer || []; 673 function gtag(){dataLayer.push(arguments);} 674 gtag('js', new Date()); 675 676 gtag('config', ""); 677</script>
677 678 679
680<script> 681 $(document).ready(function () { 682 683 var gsDataBaseUrl = 'https://cdn.jsdelivr.net/gh/ximinng.github.io@' 684 685 $.getJSON(gsDataBaseUrl + "google-scholar-stats/gs_data.json", function (data) { 686 var totalCitation = data['citedby'] 687 document.getElementById('total_cit').innerHTML = totalCitation; 688 var citationEles = document.getElementsByClassName('show_paper_citations') 689 Array.prototype.forEach.call(citationEles, element => { 690 var paperId = element.getAttribute('data') 691 var numCitations = data['publications'][paperId]['num_citations'] 692 element.innerHTML = '| Citations: ' + numCitations; 693 }); 694 }); 695 }) 696</script>
696 697 698 699 </body> 700</html>
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.