1<!DOCTYPE html> 2<html lang="en"> 3 4<head> 5 <meta charset="utf-8"> 6 <meta name="viewport" content="width=device-width, initial-scale=1"> 7 8 <!-- Primary Meta Tags --> 9 <meta name="title" 10 content="EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric Video Understanding and Generation"> 11 <meta name="description" 12 content="EmoVid is a multimodal emotion-centric video dataset for understanding and generation. The project page hosts the paper, appendix, figures, and interactive visualizations."> 13 <meta name="keywords" 14 content="EmoVid, emotion, video, dataset, multimodal, video understanding, video generation, affective computing, AAAI 2026"> 15 <meta name="author" content="EmoVid Authors"> 16 <meta name="robots" content="index, follow"> 17 <meta name="language" content="English"> 18 19 <!-- Open Graph / Facebook --> 20 <meta property="og:type" content="article"> 21 <meta property="og:site_name" content="EmoVid"> 22 <meta property="og:title" 23 content="EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric Video Understanding and Generation"> 24 <meta property="og:description" 25 content="A multimodal emotion-centric video dataset for understanding and generation. AAAI 2026 Oral."> 26 <meta property="og:url" content=""> 27 <meta property="og:image" content="Materials/Images/headimage.png"> 28 <meta property="og:image:width" content="1200"> 29 <meta property="og:image:height" content="630"> 30 <meta property="og:image:alt" content="EmoVid research preview image"> 31 <meta property="article:published_time" content="2026-01-01T00:00:00.000Z"> 32 <meta property="article:author" content="EmoVid Authors"> 33 <meta property="article:section" content="Research"> 34 <meta property="article:tag" content="emotion"> 35 <meta property="article:tag" content="video"> 36 37 <!-- Twitter --> 38 <meta name="twitter:card" content="summary_large_image"> 39 <meta name="twitter:site" content=""> 40 <meta name="twitter:creator" content=""> 41 <meta name="twitter:title" 42 content="EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric Video Understanding and Generation"> 43 <meta name="twitter:description" 44 content="A multimodal emotion-centric video dataset for understanding and generation. AAAI 2026 Oral."> 45 <meta name="twitter:image" content="Materials/Images/headimage.png"> 46 <meta name="twitter:image:alt" content="EmoVid research preview image"> 47 48 <!-- Academic/Research Specific --> 49 <meta name="citation_title" 50 content="EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric Video Understanding and Generation"> 51 <meta name="citation_author" content="EmoVid Authors"> 52 <meta name="citation_publication_date" content="2026"> 53 <meta name="citation_conference_title" content="AAAI"> 54 <meta name="citation_pdf_url" content="Materials/EmoVid.pdf"> 55 56 <!-- Additional SEO --> 57 <meta name="theme-color" content="#2563eb"> 58 <meta name="msapplication-TileColor" content="#2563eb"> 59 <meta name="apple-mobile-web-app-capable" content="yes"> 60 <meta name="apple-mobile-web-app-status-bar-style" content="default"> 61 62 <!-- Preconnect for performance --> 63 <link rel="preconnect" href="https://fonts.googleapis.com"> 64 <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin> 65 <link rel="preconnect" href="https://ajax.googleapis.com"> 66 <link rel="preconnect" href="https://documentcloud.adobe.com"> 67 <link rel="preconnect" href="https://cdn.jsdelivr.net"> 68 69 <title>EmoVid: A Multimodal Emotion Video Dataset | AAAI 2026</title> 70 71 <!-- Favicon and App Icons --> 72 <link rel="icon" type="image/x-icon" href="static/images/logo.ico"> 73 <link rel="apple-touch-icon" href="static/images/logo.ico"> 74 75 <!-- Critical CSS - Load synchronously --> 76 <link rel="stylesheet" href="static/css/bulma.min.css"> 77 <link rel="stylesheet" href="static/css/index.css"> 78 79 <!-- Non-critical CSS - Load asynchronously --> 80 <link rel="preload" href="static/css/bulma-carousel.min.css" as="style" 81 onload="this.onload=null;this.rel='stylesheet'"> 82 <link rel="preload" href="static/css/bulma-slider.min.css" as="style" onload="this.onload=null;this.rel='stylesheet'"> 83 <link rel="preload" href="static/css/fontawesome.all.min.css" as="style" 84 onload="this.onload=null;this.rel='stylesheet'"> 85 <link rel="preload" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css" as="style" 86 onload="this.onload=null;this.rel='stylesheet'"> 87 88 <!-- Fallback for browsers that don't support preload --> 89 <noscript> 90 <link rel="stylesheet" href="static/css/bulma-carousel.min.css"> 91 <link rel="stylesheet" href="static/css/bulma-slider.min.css"> 92 <link rel="stylesheet" href="static/css/fontawesome.all.min.css"> 93 <link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css"> 94 </noscript> 95 96 <!-- Fonts - Optimized loading --> 97 <link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap" rel="stylesheet"> 98 99 <style> 100 .eviz-wrapper { 101 background: rgba(255, 255, 255, 0.92); 102 border-radius: 16px; 103 padding: 2rem; 104 box-shadow: 0 15px 35px rgba(0, 0, 0, 0.08); 105 } 106 107 .eviz-header { 108 text-align: center; 109 margin-bottom: 2rem; 110 } 111 112 .eviz-section { 113 margin-bottom: 2.5rem; 114 } 115 116 .eviz-section:last-child { 117 margin-bottom: 0; 118 } 119 120 .eviz-section-title { 121 font-weight: 700; 122 margin-bottom: 0.5rem; 123 font-size: 1.75rem; 124 text-align: center; 125 } 126 127 .eviz-section-subtitle { 128 margin-top: -0.5rem; 129 margin-bottom: 1.5rem; 130 color: #6b7280; 131 font-size: 1rem; 132 text-align: center; 133 } 134 135 /* Dataset Sample Styles */ 136 .eviz-dataset-grid { 137 display: grid; 138 grid-template-columns: repeat(auto-fit, minmax(300px, 1fr)); 139 gap: 1.5rem; 140 } 141 142 .eviz-dataset-emotion { 143 background: #ffffff; 144 border-radius: 12px; 145 padding: 1.25rem; 146 box-shadow: 0 8px 20px rgba(0, 0, 0, 0.06); 147 transition: transform 0.2s ease, box-shadow 0.2s ease; 148 } 149 150 .eviz-dataset-emotion:hover { 151 transform: translateY(-5px); 152 box-shadow: 0 12px 25px rgba(0, 0, 0, 0.08); 153 } 154 155 .eviz-dataset-emotion h3 { 156 text-align: left; 157 text-transform: capitalize; 158 font-weight: 700; 159 margin-bottom: 1rem; 160 font-size: 1.25rem; 161 } 162 163 .eviz-dataset-categories { 164 display: grid; 165 grid-template-columns: repeat(3, 1fr); 166 gap: 1rem; 167 } 168 169 .eviz-dataset-category h5 { 170 font-size: 0.75rem; 171 text-transform: uppercase; 172 letter-spacing: 0.05em; 173 color: #6b7280; 174 margin-bottom: 0.5rem; 175 } 176 177 .eviz-dataset-category video { 178 height: 120px; 179 object-fit: cover; 180 } 181 182 .eviz-dataset-category video { 183 height: 120px; 184 object-fit: cover; 185 } 186 187 /* Custom Roles Styles */ 188 .eviz-custom-grid { 189 display: grid; 190 grid-template-columns: repeat(auto-fit, minmax(300px, 1fr)); 191 gap: 1.5rem; 192 } 193 194 .eviz-role { 195 background: #ffffff; 196 border-radius: 12px; 197 padding: 1rem; 198 box-shadow: 0 8px 20px rgba(0, 0, 0, 0.06); 199 display: flex; 200 flex-direction: column; 201 align-items: center; 202 transition: transform 0.2s ease, box-shadow 0.2s ease; 203 } 204 205 .eviz-role:hover { 206 transform: translateY(-5px); 207 box-shadow: 0 12px 25px rgba(0, 0, 0, 0.08); 208 } 209 210 .eviz-role h3 { 211 text-transform: capitalize; 212 font-weight: 700; 213 text-align: center; 214 margin-bottom: 0.75rem; 215 font-size: 1.1rem; 216 } 217 218 .eviz-role h3 span { 219 font-size: 0.8rem; 220 display: block; 221 color: #5c5f7a; 222 font-weight: 500; 223 margin-top: 0.2rem; 224 } 225 226 .eviz-role-note { 227 display: flex; 228 align-items: center; 229 justify-content: center; 230 text-align: center; 231 padding: 0.5rem; 232 border-radius: 10px; 233 background: rgba(102, 126, 234, 0.1); 234 color: #4338ca; 235 font-weight: 500; 236 margin-bottom: 1rem; 237 font-size: 0.8rem; 238 width: 100px; 239 height: 100px; 240 box-shadow: 0 6px 16px rgba(0, 0, 0, 0.1); 241 } 242 243 .eviz-input-image { 244 width: 100px; 245 height: 100px; 246 object-fit: cover; 247 border-radius: 10px; 248 margin-bottom: 1rem; 249 box-shadow: 0 6px 16px rgba(0, 0, 0, 0.1); 250 } 251 252 .eviz-emotions-grid { 253 display: grid; 254 grid-template-columns: repeat(4, 1fr); 255 gap: 0.5rem; 256 width: 100%; 257 } 258 259 .eviz-emotion-video h4 { 260 font-size: 0.7rem; 261 margin-bottom: 0.25rem; 262 text-transform: capitalize; 263 color: #6b7280; 264 font-weight: 500; 265 } 266 267 .eviz-emotion-video video, 268 .eviz-dataset-category video { 269 width: 100%; 270 border-radius: 8px; 271 box-shadow: 0 4px 12px rgba(0, 0, 0, 0.08); 272 background-color: #f0f2f5; 273 } 274 275 @media (max-width: 768px) { 276 .eviz-wrapper { 277 padding: 1.5rem; 278 } 279 280 .eviz-dataset-categories { 281 grid-template-columns: 1fr; 282 } 283 284 .eviz-emotions-grid { 285 grid-template-columns: repeat(2, 1fr); 286 } 287 } 288 </style> 289 290 <!-- Defer non-critical JavaScript --> 291
291<script defer src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
291 292
292<script defer src="https://documentcloud.adobe.com/view-sdk/main.js"></script>
292 293
293<script defer src="static/js/fontawesome.all.min.js"></script>
293 294
294<script defer src="static/js/bulma-carousel.min.js"></script>
294 295
295<script defer src="static/js/bulma-slider.min.js"></script>
295 296
296<script defer src="static/js/index.js"></script>
296 297 298 <!-- Structured Data for Academic Papers --> 299
299<script type="application/ld+json"> 300 { 301 "@context": "https://schema.org", 302 "@type": "ScholarlyArticle", 303 "headline": "EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric Video Understanding and Generation", 304 "description": "EmoVid is a multimodal emotion-centric video dataset for understanding and generation. The project page hosts the paper, appendix, figures, and interactive visualizations.", 305 "author": [ 306 { 307 "@type": "Person", 308 "name": "EmoVid Authors", 309 "affiliation": { 310 "@type": "Organization", 311 "name": "" 312 } 313 } 314 ], 315 "datePublished": "2026-01-01", 316 "publisher": { 317 "@type": "Organization", 318 "name": "AAAI" 319 }, 320 "url": "", 321 "image": "Materials/Images/headimage.png", 322 "keywords": ["EmoVid", "emotion", "video", "dataset", "multimodal", "video generation", "affective computing"], 323 "abstract": "", 324 "citation": "", 325 "isAccessibleForFree": true, 326 "license": "https://creativecommons.org/licenses/by/4.0/", 327 "mainEntity": { 328 "@type": "WebPage", 329 "@id": "" 330 }, 331 "about": [ 332 { 333 "@type": "Thing", 334 "name": "Emotion-Centric Video Understanding" 335 }, 336 { 337 "@type": "Thing", 338 "name": "Video Generation" 339 } 340 ] 341 } 342 </script>
342 343 344 <!-- Website/Organization Structured Data --> 345
345<script type="application/ld+json"> 346 { 347 "@context": "https://schema.org", 348 "@type": "Organization", 349 "name": "EmoVid", 350 "url": "", 351 "logo": "static/images/favicon.ico", 352 "sameAs": [] 353 } 354 </script>
354 355</head> 356 357<body> 358 359 <!-- Scroll to Top Button --> 360 <button class="scroll-to-top" onclick="scrollToTop()" title="Scroll to top" aria-label="Scroll to top"> 361 <i class="fas fa-chevron-up"></i> 362 </button> 363 364 <!-- More Works Dropdown --> 365 <!-- <div class="more-works-container"> 366 <button class="more-works-btn" onclick="toggleMoreWorks()" title="View More Works from Our Lab"> 367 <i class="fas fa-flask"></i> 368 More Works 369 <i class="fas fa-chevron-down dropdown-arrow"></i> 370 </button> 371 <div class="more-works-dropdown" id="moreWorksDropdown"> 372 <div class="dropdown-header"> 373 <h4>More Works from Our Lab</h4> 374 <button class="close-btn" onclick="toggleMoreWorks()"> 375 <i class="fas fa-times"></i> 376 </button> 377 </div> 378 <div class="works-list"> 379 <a href="#" class="work-item" target="_blank"> 380 <div class="work-info"> 381 <h5>Paper Title 1</h5> 382 <p>Brief description of the work and its main contribution.</p> 383 <span class="work-venue">Conference/Journal</span> 384 </div> 385 <i class="fas fa-external-link-alt"></i> 386 </a> 387 <a href="#" class="work-item" target="_blank"> 388 <div class="work-info"> 389 <h5>Paper Title 2</h5> 390 <p>Brief description of the work and its main contribution.</p> 391 <span class="work-venue">Conference/Journal</span> 392 </div> 393 <i class="fas fa-external-link-alt"></i> 394 </a> 395 </div> 396 </div> 397 </div> --> 398 399 <main id="main-content"> 400 <section class="hero"> 401 <div class="hero-body"> 402 <div class="container is-max-desktop"> 403 <div class="columns is-centered"> 404 <div class="column has-text-centered"> 405 <h1 class="title is-1 publication-title">EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric 406 Video Understanding and Generation</h1> 407 <div class="is-size-5 publication-authors" style="margin-top:0.5rem;"> 408 <span class="author-block"><a href="https://zane-zyqiu.github.io" target="_blank" 409 rel="noopener">Zongyang Qiu</a><sup>*</sup>,</span> 410 <span class="author-block"><a href="https://cislab.hkust-gz.edu.cn/members/bingyuan-wang" 411 target="_blank" rel="noopener">Bingyuan Wang</a><sup>*</sup>,</span> 412 <span class="author-block"><a href="https://github.com/May-apple" target="_blank" rel="noopener">Xingbei 413 Chen</a>,</span> 414 <span class="author-block"><a href="https://yingqinghe.github.io" target="_blank" 415 rel="noopener">Yingqing He</a>,</span> 416 <span class="author-block"><a href="https://cislab.hkust-gz.edu.cn/members/zeyu-wang" target="_blank" 417 rel="noopener">Zeyu Wang</a><sup>â </sup></span> 418 </div> 419 420 <div class="is-size-5 publication-authors" style="margin-top:0.25rem;"> 421 <span class="author-block">HKUST(GZ)</span> 422 <span class="author-block">Fudan University</span> 423 <span class="author-block">HKUST</span><br> 424 <span class="author-block">AAAI 2026 Oral</span> 425 <span class="eql-cntrb"><small><br><sup>*</sup>Indicates Equal Contribution, <sup>â </sup>Corresponding Author</small></span> 426 </div> 427 428 <div class="column has-text-centered" style="margin-top:0.75rem;"> 429 <div class="publication-links"> 430 <span class="link-block"> 431 <a href="https://www.arxiv.org/abs/2511.11002" target="_blank" rel="noopener" 432 class="external-link button is-normal is-rounded is-dark"> 433 <span class="icon"> 434 <i class="fas fa-file-pdf"></i> 435 </span> 436 <span>Paper</span> 437 </a> 438 </span> 439 440 <span class="link-block"> 441 <a href="https://docs.google.com/forms/d/e/1FAIpQLSekp08DeWKxSbwEMkor1wIyubcqFfUwTEvCVpTlW1pga3pqaA/viewform?usp=dialog" 442 target="_blank" 443 rel="noopener" 444 class="external-link button is-normal is-rounded is-dark">
445 <span class="icon"> 446 <i class="fas fa-database"></i> 447 </span> 448 <span>Dataset</span> 449 </a> 450 </span> 451 452 <!-- <span class="link-block"> 453 <a class="external-link button is-normal is-rounded is-light" aria-disabled="true"> 454 <span class="icon"> 455 <i class="fab fa-github"></i> 456 </span> 457 <span>Code (coming soon)</span> 458 </a> 459 </span> --> 460 </div> 461 </div> 462 </div> 463 </div> 464 </div> 465 </div> 466 </section> 467 468 <!-- Teaser image / banner --> 469 <section class="hero teaser"> 470 <div class="container is-max-desktop"> 471 <div class="hero-body has-text-centered"> 472 <img src="Materials/Images/headimage.png" alt="EmoVid overview" 473 style="max-width:100%; border-radius:8px; box-shadow:0 10px 25px rgba(0,0,0,0.08);" loading="eager" /> 474 <h2 class="subtitle has-text-centered" style="margin-top:1rem;"> 475 EmoVid provides a multimodal, emotion-centric video benchmark for both understanding and generation tasks. 476 </h2> 477 </div> 478 </div> 479 </section> 480 <!-- End teaser image --> 481 482 <!-- Paper abstract --> 483 <section class="section hero is-light"> 484 <div class="container is-max-desktop"> 485 <div class="columns is-centered has-text-centered"> 486 <div class="column is-four-fifths"> 487 <h2 class="title is-3">Abstract</h2> 488 <div class="content has-text-justified"> 489 <p> 490 We introduce <strong>EmoVid</strong>, the first multimodal, emotion-annotated video dataset specifically 491 designed for artistic media, which includes cartoon animations, movie clips, and animated stickers. Each 492 video is annotated with emotion labels, visual attributes (brightness, colorfulness, hue), and text 493 captions. Through systematic analysis, we uncover spatial and temporal patterns linking visual features 494 to emotional perceptions across diverse video forms. Building on these insights, we develop an 495 emotion-conditioned video generation technique by fine-tuning the Wan2.1 model. The results show a 496 significant improvement in both quantitative metrics and the visual quality of generated videos for 497 text-to-video and image-to-video tasks. 498 </p> 499 </div> 500 </div> 501 </div> 502 </div> 503 </section> 504 <!-- End paper abstract --> 505 506 <!-- Generation and dataset visualizations --> 507 <section class="section" id="visualizations"> 508 <div class="container is-max-desktop"> 509 <div class="eviz-wrapper"> 510 <div class="eviz-header"> 511 <h2 class="title is-3">Generation and Dataset Visualizations</h2> 512 <p class="subtitle is-5">Curated multimodal samples from EmoVid, combining representative dataset snippets 513 and controllable custom-role generations.</p> 514 </div> 515 516 <div class="eviz-section"> 517 <h3 class="eviz-section-title">ð Dataset Sample</h3> 518 <p class="eviz-section-subtitle">Tri-modal snippets (animation, movie, sticker) for every emotion class. 519 </p> 520 <div class="eviz-dataset-grid" id="eviz-dataset-content"></div> 521 </div> 522 523 <div class="eviz-section"> 524 <h3 class="eviz-section-title">ð Custom Roles</h3> 525 <p class="eviz-section-subtitle">Role-conditioned generations showcasing controllable styles.</p> 526 <div class="eviz-custom-grid" id="eviz-custom-content"></div> 527 </div> 528 529 </div> 530 </div> 531 </section> 532 <!-- End Generation and dataset visualizations --> 533 534 <!-- Key Findings --> 535 <section class="section"> 536 <div class="container is-max-desktop"> 537 <h2 class="title is-3 has-text-centered">Key Findings From EmoVid</h2> 538 <div class="columns is-multiline is-centered is-variable is-4"> 539 540 <div class="column is-half"> 541 <div class="content"> 542 <figure class="image"> 543 <img src="Materials/Images/color attributes.png" alt="Color and brightness attributes per emotion" 544 style="border-radius:8px; box-shadow: 0 8px 16px rgba(0,0,0,0.1);"> 545 </figure> 546 <h4 class="title is-6 has-text-centered" style="margin-top: 1rem; margin-bottom: 0.5rem;">Color and 547 Brightness are Strong Emotional Indicators</h4> 548 <p class="has-text-justified"> 549 Our analysis of the EmoVid dataset reveals a clear and quantifiable link between low-level visual 550 attributes and high-level emotional categories. We found that positive-valence emotions (like 551 <strong>contentment</strong> and <strong>amusement</strong>) are generally brighter and more colorful. 552 Conversely, high-arousal emotions (like <strong>anger</strong> and <strong>excitement</strong>) also 553 trend towards higher colorfulness but with lower brightness. This spatial layout in the color space 554 supports the robustness of the valence-arousal model. 555 </p> 556 </div> 557 </div> 558 559 <div class="column is-half"> 560 <div class="content"> 561 <figure class="image"> 562 <img src="Materials/Images/user_study.png" alt="User study results" 563 style="border-radius:8px; box-shadow: 0 8px 16px rgba(0,0,0,0.1);"> 564 </figure> 565 <h4 class="title is-6 has-text-centered" style="margin-top: 1rem; margin-bottom: 0.5rem;">Fine-Tuning 566 on EmoVid Decisively Wins User Preference</h4> 567 <p class="has-text-justified"> 568 To validate the practical value of our dataset, we fine-tuned the Wan2.1 generative model on EmoVid. In 569 a controlled perceptual user study, our fine-tuned model was overwhelmingly preferred over strong 570 baselines (Wan-Original and CogVideoX). For <strong>Emotion Expression</strong>, our model was ranked 571 first in 66.2% of comparisons. For <strong>Aesthetic Quality</strong>, it achieved a 57.9% Top-1 572 preference rate. This demonstrates EmoVid's clear utility in enhancing the emotional expressiveness of 573 generative models. 574 </p> 575 </div> 576 </div> 577 578 <div class="column is-half"> 579 <div class="content"> 580 <figure class="image"> 581 <img src="Materials/Images/emo_transition_prob.png" alt="Emotion transition matrix" 582 style="border-radius:8px; box-shadow: 0 8px 16px rgba(0,0,0,0.1);"> 583 </figure> 584 <h4 class="title is-6 has-text-centered" style="margin-top: 1rem; margin-bottom: 0.5rem;">Emotions Show 585 Strong Temporal Persistence</h4> 586 <p class="has-text-justified"> 587 A key advantage of a video dataset is the ability to study temporal dynamics. By analyzing consecutive 588 movie clips, we built an emotion transition matrix. The analysis shows that emotions have strong
589 self-persistence (the strong diagonal). Furthermore, transitions are far more likely to occur 590 <em>within</em> the same valence (e.g., from one negative emotion to another) than <em>across</em> 591 valences (e.g., from positive to negative). 592 </p> 593 </div> 594 </div> 595 596 <div class="column is-half"> 597 <div class="content"> 598 <figure class="image"> 599 <img src="Materials/Images/keywords.png" alt="Top-5 phrases per emotion" 600 style="border-radius:8px; box-shadow: 0 8px 16px rgba(0,0,0,0.1);"> 601 </figure> 602 <h4 class="title is-6 has-text-centered" style="margin-top: 1rem; margin-bottom: 0.5rem;">Text Captions 603 and Emotion Labels are Strongly Aligned</h4> 604 <p class="has-text-justified"> 605 EmoVid is a multimodal dataset, and our analysis confirms its cross-modal consistency. We performed 606 sentiment polarity analysis on the text captions for each video. The results show a clear pattern: 607 videos labeled with positive emotions (e.g., <strong>amusement</strong>, <strong>excitement</strong>, 608 <strong>contentment</strong>) have a significantly higher proportion of positive-sentiment captions, and 609 the same holds true for negative emotions. This confirms the strong semantic link between the textual 610 content and the perceived emotion. 611 </p> 612 </div> 613 </div> 614 615 </div> 616 </div> 617 </section> 618 <!-- End Key Findings --> 619 620 <!--BibTex citation --> 621 <section class="section" id="BibTeX"> 622 <div class="container is-max-desktop content"> 623 <div class="bibtex-header"> 624 <h2 class="title">BibTeX</h2> 625 <button class="copy-bibtex-btn" onclick="copyBibTeX()" title="Copy BibTeX to clipboard"> 626 <i class="fas fa-copy"></i> 627 <span class="copy-text">Copy</span> 628 </button> 629 </div> 630 <pre id="bibtex-code"><code>@misc{qiu2025emovidmultimodalemotionvideo, 631 title={EmoVid: A Multimodal Emotion Video Dataset for Emotion-Centric Video Understanding and Generation}, 632 author={Zongyang Qiu and Bingyuan Wang and Xingbei Chen and Yingqing He and Zeyu Wang}, 633 year={2025}, 634 eprint={2511.11002}, 635 archivePrefix={arXiv}, 636 primaryClass={cs.CV}, 637 url={https://arxiv.org/abs/2511.11002}, 638}</code></pre> 639 </div> 640 </section> 641 <!--End BibTex citation --> 642 </main> 643 644
644<script> 645 (function () { 646 const emotions = ['contentment', 'awe', 'amusement', 'excitement', 'sadness', 'disgust', 'fear', 'anger']; 647 const categories = ['animation', 'movie', 'sticker']; 648 const customRoles = [ 649 { 650 name: 'chiikawa', 651 displayName: 'Chiikawa I2V', 652 folder: 'chiikawa_i2v', 653 hasInputImage: true, 654 imageFile: 'input_image.png', 655 videoPrefix: 'chiikawa_', 656 subtitle: 'Image-to-Video' 657 }, 658 { 659 name: 'labubu', 660 displayName: 'Labubu I2V', 661 folder: 'labubu_i2v', 662 hasInputImage: true, 663 imageFile: 'input_image.jpg', 664 videoPrefix: 'labubu_', 665 subtitle: 'Image-to-Video' 666 }, 667 { 668 name: 'ishowspeed', 669 displayName: 'iShowSpeed I2V', 670 folder: 'ishowspeed_i2v', 671 hasInputImage: true, 672 imageFile: 'input_image.jpg', 673 videoPrefix: 'speed_', 674 subtitle: 'Image-to-Video' 675 }, 676 { 677 name: 'ghibli', 678 displayName: 'Ghibli T2V', 679 folder: 'ghibli_t2v', 680 hasInputImage: false, 681 ghibliText: 'Studio Ghibli style prompt', 682 videoPrefix: 'ghibli_', 683 subtitle: 'Text-to-Video' 684 } 685 ]; 686 687 const sampleVideos = { 688 contentment: { 689 animation: ['Aladdin_scene_1131_instance_0_1.mp4', 'How to Keep a Mummy_scene_308_instance_1_1.mp4'], 690 movie: ['-mexzYsMSro_clip_010.mp4', '0nfoP3bmd1c_clip_024.mp4'], 691 sticker: ['13047851479713149078.mp4', '4165678996664452379.mp4'] 692 }, 693 awe: { 694 animation: ['How to Keep a Mummy_scene_369_instance_0_1.mp4', 'ID:INVADED_scene_251_instance_1_1.mp4'], 695 movie: ['-c9-2KcnLXI_clip_004.mp4', '0mRRULBvuj0_clip_010.mp4'], 696 sticker: ['1151841417982281027.mp4', '693624055024178899.mp4'] 697 }, 698 amusement: { 699 animation: ['Snow White and the Seven Dwarfs_scene_482_instance_0_1.mp4', 'jiujijinghua_07_scene_298_instance_0_1.mp4'], 700 movie: ['DUMW--Zlrcs_clip_033.mp4', 'gnbjy2tWfd4_clip_012.mp4'], 701 sticker: ['1014953795366610078.mp4', '7313602290956595969.mp4'] 702 }, 703 excitement: { 704 animation: ['The Little Mermaid_scene_253_instance_0_1.mp4', 'The Tigger Movie_scene_646_instance_0_2.mp4'], 705 movie: ['3Sz7dbX2kTY_clip_017.mp4', '_KC5AJdRgBs_clip_045.mp4'], 706 sticker: ['155648567990388098.mp4', '1779437211223867166.mp4'] 707 }, 708 sadness: { 709 animation: ['How to Keep a Mummy_scene_208_instance_0_1.mp4', 'My Youth Romantic Comedy Is Wrong, as I Expected_scene_307_instance_0_1.mp4'], 710 movie: ['0tRSAQUheo0_clip_021.mp4', '3sP7UMxhGYw_clip_096.mp4'], 711 sticker: ['123466421705187905.mp4', '12414244761973568755.mp4'] 712 }, 713 disgust: { 714 animation: ['Snow White and the Seven Dwarfs_scene_68_instance_0_1.mp4', 'Tower of God_scene_141_instance_0_1.mp4'], 715 movie: ['3JLDAyAEaqI_clip_031.mp4', '7kihC0VFaQE_clip_038.mp4'], 716 sticker: ['10355761199465314392.mp4', '1538353539352474088.mp4'] 717 }, 718 fear: { 719 animation: ['Atlantis: The Lost Empire_scene_1353_instance_1_1.mp4', 'How to Keep a Mummy_scene_207_instance_1_1.mp4'], 720 movie: ['0-81bpcuz44_clip_042.mp4', '0p9Q6tDJJ1w_clip_008.mp4'], 721 sticker: ['2518671278090971395.mp4', '909784035520128291.mp4'] 722 }, 723 anger: { 724 animation: ['Aladdin_scene_1351_instance_0_1.mp4', "Xam'd: Lost Memories_scene_235_instance_0_1.mp4"], 725 movie: ['-oPHsR72Lfo_clip_045.mp4', '_nSQhWq7etg_clip_020.mp4'], 726 sticker: ['4367649449949661050.mp4', '7666337384347304302.mp4'] 727 } 728 }; 729 730 const customContainer = document.getElementById('eviz-custom-content'); 731 const datasetContainer = document.getElementById('eviz-dataset-content'); 732 733 function src(path) { 734 return encodeURI(path); 735 } 736 737 function renderCustom() { 738 if (!customContainer) return; 739 let html = '';
740 customRoles.forEach(role => { 741 const imageHtml = role.hasInputImage 742 ? `<img src="${src('Materials/generation results/custom roles/' + role.folder + '/' + role.imageFile)}" alt="${role.displayName} reference" class="eviz-input-image" loading="lazy">` 743 : `<div class="eviz-role-note">${role.ghibliText}</div>`; 744 745 html += ` 746 <div class="eviz-role"> 747 <h3>${role.displayName}<span>${role.subtitle}</span></h3> 748 ${imageHtml} 749 <div class="eviz-emotions-grid"> 750 `; 751 752 emotions.forEach(emotion => { 753 const videoSrc = src(`Materials/generation results/custom roles/${role.folder}/${role.videoPrefix}${emotion}.mp4`); 754 html += ` 755 <div class="eviz-emotion-video"> 756 <h4>${emotion}</h4> 757 <video autoplay muted loop playsinline preload="auto" loading="lazy"> 758 <source src="${videoSrc}" type="video/mp4"> 759 </video> 760 </div> 761 `; 762 }); 763 764 html += ` 765 </div> 766 </div> 767 `; 768 }); 769 770 customContainer.innerHTML = html; 771 } 772 773 function renderDataset() { 774 if (!datasetContainer) return; 775 let html = ''; 776 emotions.forEach(emotion => { 777 const entries = sampleVideos[emotion]; 778 if (!entries) return; 779 html += ` 780 <div class="eviz-dataset-emotion"> 781 <h3>${emotion}</h3> 782 <div class="eviz-dataset-categories"> 783 `; 784 785 categories.forEach(category => { 786 const files = entries[category] || []; 787 html += ` 788 <div class="eviz-dataset-category"> 789 <h5>${category}</h5> 790 <div class="category-videos"> 791 `; 792 793 files.forEach(file => { 794 const videoSrc = src(`Materials/dataset sample/${emotion}/${category}/video/${file}`); 795 html += ` 796 <video autoplay muted loop playsinline preload="auto" loading="lazy"> 797 <source src="${videoSrc}" type="video/mp4"> 798 </video> 799 `; 800 }); 801 802 html += ` 803 </div> 804 </div> 805 `; 806 }); 807 808 html += ` 809 </div> 810 </div> 811 `; 812 }); 813 datasetContainer.innerHTML = html; 814 } 815 816 renderCustom(); 817 renderDataset(); 818 })(); 819 </script>
819 820 821 <footer class="footer"> 822 <div class="container"> 823 <div class="columns is-centered"> 824 <div class="column is-8"> 825 <div class="content"> 826 <p> 827 This page was built using the <a href="https://github.com/eliahuhorwitz/Academic-project-page-template" 828 target="_blank">Academic Project Page Template</a> which was adopted from the <a 829 href="https://nerfies.github.io" target="_blank">Nerfies</a> project page. 830 You are free to borrow the source code of this website, we just ask that you link back to this page in the 831 footer. <br> This website is licensed under a <a rel="license" 832 href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative Commons 833 Attribution-ShareAlike 4.0 International License</a>. 834 </p> 835 </div> 836 </div> 837 </div> 838 </div> 839 </footer> 840 841</body> 842 843</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.