1<!DOCTYPE html> 2<html> 3<head> 4 <meta charset="utf-8"> 5 <meta name="description" 6 content="camera intrinsic setting control for text-to-image generation."> 7 <meta name="keywords" content="Camera Control, Text-to-Image Generation, Text-to-Video Generation, Diffusion Model"> 8 <meta name="viewport" content="width=device-width, initial-scale=1"> 9 <title>Generative Photography: Scene-Consistent Camera Control for Realistic Text-to-Image Synthesis</title> 10 11 <!-- Global site tag (gtag.js) - Google Analytics -->
vendor: 66 bytes, lines 11-12
11 12 <script async src="https://www.googletagmanager.com/gtag/js?id=
12G-PYVRSFMDRL
vendor: 14 bytes, lines 12-13
12"></script> 13
13<script> 14
vendor: 155 bytes, lines 14-22
14window.dataLayer = window.dataLayer || []; 15 16 function gtag() { 17 dataLayer.push(arguments); 18 } 19 20 gtag('js', new Date()); 21 22 gtag('config', '
22G-PYVRSFMDRL
vendor: 6 bytes, lines 22-23
22'); 23
23</script>
23 24 25 <link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro" 26 rel="stylesheet"> 27 28 <link rel="stylesheet" href="./static/css/bulma.min.css"> 29 <link rel="stylesheet" href="./static/css/bulma-carousel.min.css"> 30 <link rel="stylesheet" href="./static/css/bulma-slider.min.css"> 31 <link rel="stylesheet" href="./static/css/fontawesome.all.min.css"> 32 <link rel="stylesheet" 33 href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css"> 34 <link rel="stylesheet" href="./static/css/index.css"> 35 <link rel="icon" href="./static/images/favicon.svg"> 36 37
37<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
37 38
38<script defer src="./static/js/fontawesome.all.min.js"></script>
38 39
39<script src="./static/js/bulma-carousel.min.js"></script>
39 40
40<script src="./static/js/bulma-slider.min.js"></script>
40 41
41<script src="./static/js/index.js"></script>
41 42</head> 43<body> 44 45 46<section class="hero"> 47 <div class="hero-body"> 48 <div class="container is-max-desktop"> 49 <div class="columns is-centered"> 50 <div class="column has-text-centered"> 51 <h1 class="title is-1 publication-title"> 52 <span style="color:#B9770E; font-weight: bold; font-style: italic">Generative Photography</span> 53 <br> <!-- æ¢è¡ --> 54 Scene-Consistent Camera Control for 55 <br> Realistic Text-to-Image Synthesis 56 57 <br> 58 <span style="font-size: 0.6em; font-weight: normal;">CVPR 2025 Highlight & Demo</span> 59 </h1> 60<div class="is-size-5 publication-authors"> 61 <div class="author-row"> 62 <span class="author-block"> 63 <a href="https://yuyuan-space.github.io/">Yu Yuan</a><sup>1,â¡</sup>,</span> 64 <span class="author-block"> 65 <a href="https://www.linkedin.com/in/xijun-wang-747475208/">Xijun Wang</a><sup>1,â¡</sup>,</span> 66 <span class="author-block"> 67 <a href="https://shengcn.github.io/">Yichen Sheng</a><sup>2</sup>,</span> 68 <span class="author-block"> 69 <a href="https://www.linkedin.com/in/prateek-chennuri-3a25a8171/">Prateek Chennuri</a><sup>1</sup>,</span> 70 <span class="author-block"> 71 <a href="https://xg416.github.io/">Xingguang Zhang</a><sup>1</sup>,</span> 72 <span class="author-block"> 73 <a href="https://engineering.purdue.edu/ChanGroup/stanleychan.html">Stanley Chan</a><sup>1,â </sup>,</span> 74 </div> 75</div> 76 77<div class="is-size-5 publication-authors"> 78 <span class="author-block"><sup>1</sup>Purdue University,</span> 79 <span class="author-block"><sup>2</sup>NVIDIA Research</span> 80</div> 81 82<div class="is-size-6 author-footnote"> 83 <span>â¡ These authors contributed equally to the conceptualization.</span><br> 84 <span>
84â Corresponding author.</span> 85</div> 86 87 88 <div class="column has-text-centered"> 89 <div class="publication-links"> 90 <!-- PDF Link. --> 91 <span class="link-block"> 92 <a href="https://arxiv.org/abs/2412.02168" 93 class="external-link button is-normal is-rounded is-dark"> 94 <span class="icon"> 95 <i class="ai ai-arxiv"></i> 96 </span> 97 <span>arXiv</span> 98 </a> 99 </span> 100 <!-- Code Link. --> 101 <span class="link-block"> 102 <a href="https://github.com/pandayuanyu/generative-photography" 103 class="external-link button is-normal is-rounded is-dark"> 104 <span class="icon"> 105 <i class="fab fa-github"></i> 106 </span> 107 <span>Code</span> 108 </a> 109 </span> 110 <!-- Dataset Link. --> 111 <span class="link-block"> 112 <a href="https://huggingface.co/datasets/pandaphd/camera_settings" 113 class="external-link button is-normal is-rounded is-dark"> 114 <span class="icon"> 115 <img src="https://huggingface.co/front/assets/huggingface_logo-noborder.svg" alt="Hugging Face" style="width: 24px; height: 24px;"> 116 </span> 117 <span>Dataset</span> 118 </a> 119 </span> 120 <!-- HuggingFace Link. --> 121 <span class="link-block"> 122 <a href="https://huggingface.co/spaces/pandaphd/generative_photography" 123 class="external-link button is-normal is-rounded is-dark"> 124 <span class="icon"> 125 <img src="https://huggingface.co/front/assets/huggingface_logo.svg" alt="Hugging Face Logo" style="width:24px; height:24px;"> 126 </span> 127 <span>Demo</span> 128 </a> 129 </span> 130 </span> 131 </div> 132 </div> 133 </div> 134 </div> 135 </div> 136 </div> 137</section> 138<section class="hero teaser"> 139 <div class="container is-max-desktop"> 140 <div class="hero-body"> 141 <h2 class="subtitle has-text-centered"> 142 <span class="dnerf"></span> Text-to-Image Generation with an Understanding of Camera Physics! 143 </h2> 144 </div> 145 </div> 146</section> 147 148 149<style> 150 .container { 151 width: 90%; /* å 容宽度ä¸å±å¹ä¸è´ */ 152 max-width: 1440px; /* æå¤§å®½åº¦éå¶ä¸º1200pxï¼éåºå¤§å± */ 153 margin: auto; /* æ°´å¹³å± ä¸ */ 154 } 155 156 table { 157 border-collapse: collapse; 158 margin: 0 auto; 159 width: 100%; 160 max-width: 1000px; 161 } 162 163 table, th, td { 164 border: 1.5px solid rgb(230, 230, 230); 165 } 166 167 th, td { 168 text-align: center; 169 vertical-align: middle; 170 padding: 5px; 171 } 172 173 th { 174 font-weight: bold; 175 white-space: nowrap; 176 } 177 178 td:first-child { 179 width: 8%; 180 } 181 182 td:nth-child(2), td:nth-child(3), td:nth-child(4), td:nth-child(5) { 183 width: 23%; 184 } 185 186 td { 187 height: 160px; 188 } 189 190 tr:first-child th, tr:first-child td { 191 height: 30px; /* 强å¶éå¶é«åº¦ */ 192 overflow: hidden; /* 鲿¢å å®¹æº¢åº */ 193 line-height: 30px; /* åç´å± ä¸ */ 194 } 195 196 /* å¾çæ ·å¼ */ 197 .image-container img { 198 width: 130%; 199 height: auto; 200 max-width: 200px; 201 object-fit: contain; 202 } 203</style> 204 205<section> 206 <div class="container is-max-desktop"> 207 <!-- Abstract. --> 208 <div class="columns is-centered has-text-centered"> 209 <div class="column is-four-fifths"> 210 <br> 211 <br> 212 <h2 class="title is-3">Abstract</h2> 213 <div class="content has-text-justified"> 214 <p> 215 Image generation today can produce somewhat realistic images from text prompts. However, 216 if one asks the generator to synthesize a specific camera setting such as creating different fields 217 of view using a 24mm lens versus a 70mm lens, the generator will not be able to interpret and generate 218 scene-consistent images. This limitation not only hinders the adoption of generative tools in 219 professional photography but also highlights the broader challenge of al
219igning data-driven models 220 with real-world physical settings. In this paper, we introduce <strong>Generative Photography</strong>, 221 a framework that allows controlling <strong>camera intrinsic settings</strong> during content generation. 222 The core innovation of this work are the concepts of <strong>Dimensionality Lifting and 223 Differential Camera Intrinsics Learning</strong>, enabling smooth and consistent transitions 224 across different camera settings. Experimental results show that our method produces 225 significantly more scene-consistent photorealistic images than state-of-the-art models such as Stable Diffusion 3 and FLUX. 226 </p> 227 </div> 228 </div> 229 </div> 230 </div> 231</section> 232 233<section class="section"> 234 <div class="container is-max-desktop"> 235 <div class="columns is-centered has-text-centered"> 236 <div class="column is-four-fifths"> 237 <h2 class="title is-3">Motivation</h2> 238 <div class="content has-text-justified"> 239 <p> 240 Current state-of-the-art text-to-image generation models like 241 <a href="https://huggingface.co/stabilityai/stable-diffusion-3-medium">Stable Diffusion 3 (SD3) </a> 242 and <a href="https://github.com/black-forest-labs/flux"> FLUX </a> 243 face two major limitations: 244 the inability to accurately <strong>interpret</strong> camera-specific settings and 245 the challenge of maintaining <strong>consistency</strong> in the base scene (pls see below examples). 246 247 This work introduces a novel approach that addresses these issues, 248 enabling precise camera setting control and consistent scenes in generative models. 249 </p> 250 <table style="width:100%"> 251 <tr> 252 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Type</strong></td> 253 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Bokeh Rendering</strong></td> 254 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Focal Length</strong></td> 255 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Shutter Speed</strong></td> 256 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Color Temperature</strong></td> 257 </tr> 258 <tr> 259 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Prompts</strong></td> 260 <td class="text-cell" style="text-align:center;vertical-align: middle;">"... desserts; with <strong>bokeh blur parameter [2, 6, 10, 14, 18].</strong>"</td> 261 <td class="text-cell" style="text-align:center;vertical-align: middle;">"... office; with <strong>[45, 40, 35, 30, 25]mm lens.</strong>"</td> 262 <td class="text-cell" style="text-align:center;vertical-align: middle;">"... kitchen; with <strong>shutter speed [0.2, 0.36, 0.46, 0.66, 0.85] second.</strong>"</td> 263 <td class="text-cell" style="text-align:center;vertical-align: middle;">"... living room; with <strong>temperature [3200, 3050, 3700, 4500, 9500] kelvin.</strong>"</td> 264 </tr> 265 <tr> 266 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Generated by SD3/FLUX</strong></td> 267 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/bokeh_rendering/0_desserts_SD3.gif" alt="SD3"></td> 268 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/focal_length/0_office_FLUX.gif" alt="FLUX"></td> 269 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/shutter_speed/4_kitchen_SD3.gif" alt="SD3"></td> 270 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/color_temperature/0_room_FLUX.gif" alt="FLUX"></td> 271 </tr> 272 <tr> 273 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Ours</strong></td> 274 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/bokeh_rendering/0_desserts_Our.gif" alt="Ours"></td> 275 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/focal_length/0_office_Our.gif" alt="Ours"></td> 276 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/shutter_speed/4_kitchen_Our.gif" alt="Ours"></td> 277 <td class="image-container" style="vertical-align: middle;"><img src="static/videos/color_temperature/0_room_Our.gif" alt="Ours"></td> 278 </tr> 279 </table> 280 </div> 281 </div> 282 </div> 283 </div> 284</section> 285 286 287<section class="section"> 288 <div class="container is-max-desktop"> 289 <div class="columns is-centered has-text-centered"> 290 <div class="column is-four-fifths"> 291 <h2 class="title is-3">What is Generative Photography?</h2> 292 <div class="content has-text-justified"> 293 <p> 294 Generative photography is a new paradigm in photography where content is generated instead of captured. 295 On top of an existing text-to-image generation process, 296 we demand the model to comprehend the typical camera settings: 297 adjusting the aperture, shutter speed, focal length, and color temperature. 298 A successful generative photography method should satisfy three objectives: 299 (1) the camera effects are <strong>realistically rendered</strong>; 300 (2) by changing the camera settings, the content of the scene is <strong>not altered</strong>, e.g., 301 the buildings remain the same buildings and persons remain the same persons; 302 (3) adding camera awareness does not degrade the image <strong>quality</strong> when compared 303 to the baseline models that do not possess this property. 304 </p> 305 </div> 306 </div> 307 </div> 308</section> 309 310 311 312<section class="section"> 313 <div class="container is-max-desktop"> 314 <!-- Method. --> 315 <div class="columns is-centered has-text-centered"> 316 <div class="column is-four-fifths"> 317 <h2 class="title is-3">Methods</h2> 318 <img src="./static/images/Pipeline.png" alt="framework"> 319 <div class="content has-text-justified"> 320 <p>Our high-level idea consists of two main points: 321 <li> 322 <strong>Dimensionality Lifting.</strong> 323 We should lift the space-only problem to a space-camera joint problem. 324 The lifting and the joint optimization will enable new 325 <strong>disentanglement</strong> capabilities and ensure <strong>consistency</strong>. 326 </li> 327 <li> 328 <strong>Differential Camera Intrinsics Learning.</strong> 329 A strategy designed to explicitly 330 capture differences in camera intrinsic settings, operating 331 simultaneously at both data and network architecture levels
332 to reinforce consistent scene representation. 333 </li> 334 </p> 335 </div> 336 <img src="./static/images/Data.png" alt="framework" style="width: 80%;"> 337 <div class="content has-text-justified"> 338 <p>We construct differential data for the same scene using 339 the approach illustrated above. 340 Differential data is advantageous as it enables models to 341 focus on and learn from the differences between images, 342 rather than being overwhelmed by the vast diversity of real-world data. 343 </p> 344 </div> 345 346 </div> 347 </div> 348 <!--/ Method. --> 349</section> 350 351<section class="section"> 352 <div class="container is-max-desktop"> 353 <div class="columns is-centered has-text-centered"> 354 <div class="column is-four-fifths"> 355 <h2 class="title is-3">Gallery</h2> 356 <div class="content has-text-justified"> 357 <p> 358 We showcase visual results generated by different methods on camera bokeh rendering, 359 focal length, shutter speed and color temperature control. 360 Both <a href="https://huggingface.co/stabilityai/stable-diffusion-3-medium">Stable Diffusion 3 (SD3) </a> 361 and <a href="https://github.com/black-forest-labs/flux"> FLUX </a> generate images with fixed random seed. 362 363 Both <a href="https://animatediff.github.io/">AnimateDiff </a> 364 and <a href="https://hehao13.github.io/projects-CameraCtrl/"> CameraCtrl </a> 365 have been fine-tuned/trained on our contrastive data. 366 367 </p> 368 <table> 369 <tr> 370 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Type</strong></td> 371 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Bokeh Rendering</strong></td> 372 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Focal Length</strong></td> 373 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Shutter Speed</strong></td> 374 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Color Temperature</strong></td> 375 </tr> 376 <tr> 377 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Prompts</strong></td> 378 <td class="text-cell" style="text-align:center;vertical-align: middle;">"A horse with a white face stands in a grassy field, looking at the camera; 379 with <strong>bokeh blur parameter [28.0, 14.0, 10.0, 6.0, 2.0].</strong>"</td> 380 <td class="text-cell" style="text-align:center;vertical-align: middle;">"A beautiful garden filled with red roses and green leaves 381 ; with <strong>[24.9, 36.9, 48.9, 60.9, 69.9]mm lens.</strong>"</td> 382 <td class="text-cell" style="text-align:center;vertical-align: middle;">"A blue pot with a plant in it is placed on a window sill, surrounded by other potted plants; 383 with <strong>shutter speed [0.88, 0.68, 0.48, 0.38, 0.28] second.</strong>"</td> 384 <td class="text-cell" style="text-align:center;vertical-align: middle;">"A collection of trash cans and a potted plant are seen in the image. 385 The trash cans are individually in blue, black and yellow; 386 with <strong>temperature [3100.0, 4000.0, 8000.0, 7000.0, 3000.0] kelvin.</strong>"</td> 387 </tr> 388 <tr> 389 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>SD3</strong></td> 390 <td><img src="static/videos/bokeh_rendering/1_horse_SD3.gif" alt="SD3"></td> 391 <td><img src="static/videos/focal_length/4_garden_SD3.gif" alt="SD3"></td> 392 <td><img src="static/videos/shutter_speed/3_pot_SD3.gif" alt="SD3"></td> 393 <td><img src="static/videos/color_temperature/4_trash_SD3.gif" alt="SD3"></td> 394 </tr> 395 <tr> 396 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>FLUX</strong></td> 397 <td><img src="static/videos/bokeh_rendering/1_horse_FLUX.gif" alt="FLUX"></td> 398 <td><img src="static/videos/focal_length/4_garden_FLUX.gif" alt="FLUX"></td> 399 <td><img src="static/videos/shutter_speed/3_pot_FLUX.gif" alt="FLUX"></td> 400 <td><img src="static/videos/color_temperature/4_trash_FLUX.gif" alt="FLUX"></td> 401 </tr> 402 <tr> 403 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>AnimateDiff</strong></td> 404 <td><img src="static/videos/bokeh_rendering/1_horse_AnimateDiff.gif" alt="AnimateDiff"></td> 405 <td><img src="static/videos/focal_length/4_garden_AnimateDiff.gif" alt="AnimateDiff"></td> 406 <td><img src="static/videos/shutter_speed/3_pot_AnimateDiff.gif" alt="AnimateDiff"></td> 407 <td><img src="static/videos/color_temperature/4_trash_AnimateDiff.gif" alt="AnimateDiff"></td> 408 </tr> 409 <tr> 410 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>CameraCtrl</strong></td> 411 <td><img src="static/videos/bokeh_rendering/1_horse_CameraCtrl.gif" alt="CameraCtrl"></td> 412 <td><img src="static/videos/focal_length/4_garden_CameraCtrl.gif" alt="CameraCtrl"></td> 413 <td><img src="static/videos/shutter_speed/3_pot_CameraCtrl.gif" alt="CameraCtrl"></td> 414 <td><img src="static/videos/color_temperature/4_trash_CameraCtrl.gif" alt="CameraCtrl"></td> 415 </tr> 416 <tr> 417 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Ours</strong></td> 418 <td><img src="static/videos/bokeh_rendering/1_horse_Our.gif" alt="Ours"></td> 419 <td><img src="static/videos/focal_length/4_garden_Our.gif" alt="Ours"></td> 420 <td><img src="static/videos/shutter_speed/3_pot_Our.gif" alt="Ours"></td> 421 <td><img src="static/videos/color_temperature/4_trash_Our.gif" alt="Ours"></td> 422 </tr> 423 424 425 426 427 </table> 428 <br> 429 <p>We show more camera-controlled videos generated by our method.</p> 430 <table> 431 <tr> 432 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Bokeh Rendering</strong></td> 433 <td><img src="static/videos/bokeh_rendering/2_rose_Our.gif" alt="Our bokeh rendering"></td> 434 <td><img src="static/videos/bokeh_rendering/3_plants_Our.gif" alt="Our bokeh rendering"></td> 435 <td><img src="static/videos/bokeh_rendering/4_backpack_Our.gif" alt="Our bokeh rendering"></td> 436 </tr> 437 <tr> 438 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Prompts</strong></td> 439 <td colspan="1"><center>"A single pink rose stands out in a field of green grass; 440 with <strong>bokeh blur parameter [17.0, 14.0, 11.0, 8.0, 3.0].</strong>"</center></td> 441 <td colspan="1"><center>"A variety of potted plants are displayed on a window sill, 442 with some of them placed in yellow and white cups; 443 with <strong>
443bokeh blur parameter [18.0, 14.0, 10.0, 6.0, 2.0].</strong>"</center></td> 444 <td colspan="1"><center>"A colorful backpack with a floral pattern is 445 sitting on a table next to a computer monitor; 446 with <strong>bokeh blur parameter [2.0, 8.0, 13.0, 20.0, 29.0].</strong> 447 "</center></td> 448 </tr> 449 </table> 450 <table> 451 <tr> 452 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Focal Length</strong></td> 453 <td><img src="static/videos/focal_length/1_beach_Our.gif" alt="Our focal length"></td> 454 <td><img src="static/videos/focal_length/2_trail_Our.gif" alt="Our focal length"></td> 455 <td><img src="static/videos/focal_length/3_lake_Our.gif" alt="Our focal length"></td> 456 </tr> 457 <tr> 458 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Prompts</strong></td> 459 <td colspan="1"><center>"A clean beach with waves and a few footprints; 460 with <strong>[31.0, 36.0, 42.0, 48.0, 54.0]mm lens.</strong>"</center></td> 461 <td colspan="1"><center>"A quiet mountain trail with rocks and pine trees lining the path; 462 with <strong>[25.0, 33.0, 44.0, 58.0, 68.0]mm lens.</strong>"</center></td> 463 <td colspan="1"><center>"A peaceful lake surrounded by tall grass and small rocks; 464 with <strong>[30.0, 35.0, 45.0, 60.0, 70.0]mm lens.</strong>"</center></td> 465 </tr> 466 </table> 467 <table> 468 <tr> 469 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Shutter Speed</strong></td> 470 <td><img src="static/videos/shutter_speed/0_sofa_Our.gif" alt="Our shutter speed"></td> 471 <td><img src="static/videos/shutter_speed/1_horses_Our.gif" alt="Our shutter speed"></td> 472 <td><img src="static/videos/shutter_speed/2_table_Our.gif" alt="Our shutter speed"></td> 473 </tr> 474 <tr> 475 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Prompts</strong></td> 476 <td colspan="1"><center>"A cozy living room with a large, comfy sofa and a coffee table; 477 with <strong>shutter speed [0.2, 0.3, 0.4, 0.5, 0.6] second.</strong>"</center></td> 478 <td colspan="1"><center>"A group of horses graze on a grassy field near a lake; 479 with <strong>shutter speed [0.77, 0.64, 0.43, 0.29, 0.19] second.</strong>"</center></td> 480 <td colspan="1"><center>"A spacious dining room with a wooden table and chairs around it; 481 with <strong>shutter speed [0.89, 0.64, 0.49, 0.29, 0.19] second.</strong>"</center></td> 482 </tr> 483 </table> 484 <table> 485 <tr> 486 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Color Temp.</strong></td> 487 <td><img src="static/videos/color_temperature/1_conference_Our.gif" alt="Our color temperature"></td> 488 <td><img src="static/videos/color_temperature/2_house_Our.gif" alt="Our color temperature"></td> 489 <td><img src="static/videos/color_temperature/3_castle_Our.gif" alt="Our color temperature"></td> 490 </tr> 491 <tr> 492 <td class="text-cell" style="text-align:center;vertical-align: middle;"><strong>Prompts</strong></td> 493 <td colspan="1"><center>"The image depicts a large conference room with a long table surrounded by numerous chairs; 494 with <strong>temperature [8000.0, 6000.0, 4000.0, 3000.0, 2000.0] kelvin.</strong>"</center></td> 495 <td colspan="1"><center>"A view of a house surrounded by trees, with a green lawn in front of it; 496 with <strong>temperature [3600.0, 2600.0, 4600.0, 5600.0, 6600.0] kelvin.</strong>"</center></td> 497 <td colspan="1"><center>"A beautiful view of a city with a castle and a large body of water; 498 with <strong>temperature [3000.0, 6000.0, 7000.0, 8000.0, 9000.0] kelvin.</strong>"</center></td> 499 </tr> 500 </table> 501 502 503 </div> 504 </div> 505 </div> 506 </div> 507</section> 508 509 510<section class="section" id="BibTeX">
511 <div class="container is-max-desktop content"> 512 <h2 class="title">BibTeX</h2> 513 <pre><code>@article{Yuan_2024_GenPhoto, 514 title={Generative Photography: Scene-Consistent Camera Control for Realistic Text-to-Image Synthesis}, 515 author={Yuan, Yu and Wang, Xijun and Sheng, Yichen and Chennuri, Prateek and Zhang, Xingguang and Chan, Stanley}, 516 journal={arXiv preprint arXiv: 2412.02168}, 517 year={2024} 518}</code></pre> 519 </div> 520</section> 521 522 523<footer class="footer"> 524 <div class="container"> 525 <div class="columns is-centered"> 526 <div class="column is-8"> 527 <div class="content"> 528 <p> 529 Built with <a 530 href="https://github.com/nerfies/nerfies.github.io">Nerfies</a> and <a 531 href="https://pku-yuangroup.github.io/MagicTime/">MagicTime</a> templatesâthanks for their nice jobs! 532 </p> 533 </div> 534 </div> 535 </div> 536 </div> 537</footer> 538 539</body> 540</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.