-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.html
More file actions
551 lines (495 loc) · 27.7 KB
/
index.html
File metadata and controls
551 lines (495 loc) · 27.7 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- Meta tags for social media banners, these should be filled in appropriatly as they are your "business card" -->
<!-- Replace the content tag with appropriate information -->
<meta name="description" content="DESCRIPTION META TAG">
<meta property="og:title" content="SOCIAL MEDIA TITLE TAG"/>
<meta property="og:description" content="SOCIAL MEDIA DESCRIPTION TAG TAG"/>
<meta property="og:url" content="URL OF THE WEBSITE"/>
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X630-->
<meta property="og:image" content="static/image/figure1-poster-v2_00.png" />
<meta property="og:image:width" content="1200"/>
<meta property="og:image:height" content="630"/>
<meta name="twitter:title" content="TWITTER BANNER TITLE META TAG">
<meta name="twitter:description" content="TWITTER BANNER DESCRIPTION META TAG">
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X600-->
<meta name="twitter:image" content="static/images/your_twitter_banner_image.png">
<meta name="twitter:card" content="summary_large_image">
<!-- Keywords for your paper to be indexed by-->
<meta name="keywords" content="KEYWORDS SHOULD BE PLACED HERE">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Ivy-xDetector</title>
<link rel="icon" type="image/x-icon" href="static/images/IvY-X.png">
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="static/css/bulma.min.css">
<link rel="stylesheet" href="static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="static/css/bulma-slider.min.css">
<link rel="stylesheet" href="static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="static/css/index.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="https://documentcloud.adobe.com/view-sdk/main.js"></script>
<script defer src="static/js/fontawesome.all.min.js"></script>
<script src="static/js/bulma-carousel.min.js"></script>
<script src="static/js/bulma-slider.min.js"></script>
<script src="static/js/index.js"></script>
<script src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
<script id="MathJax-script" async
src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js">
</script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<img src="static/images/IvY-X-withlabel.png" alt="Project Logo" style="max-height: 100px; margin-bottom: 1rem;">
<h1 class="title is-1 publication-title">Ivy-Fake: A Unified Explainable Framework and Benchmark for Image and Video AIGC Detection</h1>
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://github.com/arnodjiang" target="_blank">*Changjiang Jiang</a>,</span>
<span class="author-block">
<a href="https://github.com/Pi3AI" target="_blank">*Wenhui Dong</a>,</span>
<span class="author-block">
<a href="THIRD AUTHOR PERSONAL LINK" target="_blank">Zhonghao Zhang</a>,</span>
<span class="author-block">
<a href="https://chenyangsi.top" target="_blank">Chenyang Si</a>,</span>
<span class="author-block">
<a href="http://39.103.203.133/member/52" target="_blank">Fengchang Yu</a>,</span>
<span class="author-block">
<a href="https://xiaoiker.github.io/" target="_blank">Wei Peng</a>,</span>
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=Be26Qu8AAAAJ&hl=zh-CN" target="_blank">Xinbin Yuan</a>,</span>
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=On6HEc4AAAAJ&hl=zh-CN" target="_blank">Yifei Bi</a>,</span>
</span>
<span class="author-block">
<a href="https://xiaoiker.github.io/" target="_blank">Ming Zhao</a>,</span>
</span>
<span class="author-block">
<a href="https://xiaoiker.github.io/" target="_blank">Zian Zhou</a>,</span>
</span>
<span class="author-block">
<a href="https://caifeng-shan.github.io/" target="_blank">Caifeng Shan</a>,</span>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">\(\pi^3\) AI Lab</span><br/>
<sup>*</sup>Equal Contribution
</div>
<!-- Github link -->
<span class="link-block">
<a href="https://github.com/Pi3AI/Ivy-Fake" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Evaluation Code</span>
</a>
</span>
<!-- Hugging Face link -->
<span class="link-block">
<a href="https://huggingface.co/datasets/AI-Safeguard/Ivy-Fake" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<img src="https://huggingface.co/front/assets/huggingface_logo-noborder.svg" style="height: 1.1em;">
</span>
<span>Dataset</span>
</a>
</span>
<span class="link-block">
<a href="https://pi3ai.github.io/IvyFake#leaderboard" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<p style="height: 1.1em;">🏆</p>
</span>
<span>Leaderboard</span>
</a>
</span>
<!-- ArXiv abstract Link -->
<span class="link-block">
<a href="https://arxiv.org/abs/2506.00979" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
<span class="link-block">
<a href="https://huggingface.co/papers/2506.00979" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<img src="https://huggingface.co/front/assets/huggingface_logo.svg" style="height: 1.1em;">
</span>
<span>Huggingface Paper</span>
</a>
</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Teaser video-->
<section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body has-text-centered">
<img src="static/images/figure1-poster-v2_00.png" alt="Teaser Image" style="max-width: 100%; height: auto; border-radius: 12px; box-shadow: 0 4px 10px rgba(0, 0, 0, 0.1);">
<h2 class="subtitle has-text-centered" style="margin-top: 1rem;">
Overview of the IVY-FAKE framework: By conducting in-depth analysis of temporal and spatial artifacts, the framework enables explainable detection of AI-generated content
</h2>
</div>
</div>
</section>
<!-- End teaser video -->
<section class="section" id="table-of-contents">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column is-four-fifths">
<h2 class="title is-3 has-text-centered">Table of Contents</h2>
<div class="content">
<ul>
<li><a href="#overview-section">IVY-FAKE Overview</a></li>
<li><a href="#introduction-section">Introduction</a></li>
<li><a href="#dataset-section">Dataset</a></li>
<li><a href="#detector-section">Detector: IVY-XDETECTOR</a></li>
<li><a href="#experiments-section">Experiments</a></li>
<li><a href="#conclusion-section">Conclusion</a></li>
</ul>
</div>
</div>
</div>
</div>
</section>
<section class="section hero is-light" id="overview-section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">IVY-FAKE Overview</h2>
<div class="content has-text-justified">
<p>
<strong>IVY-FAKE</strong> is the first unified benchmark and explainable framework for detecting AI-generated images and videos. It provides over 100,000 annotated training samples and 5,000 evaluation examples, each paired with human-readable explanations. We introduce <strong>IVY-XDETECTOR</strong>, a vision-language model capable of multimodal AIGC detection with high accuracy and transparency. Our approach offers detailed spatial and temporal reasoning, enabling robust identification of synthetic content in real-world scenarios.
</p>
<p>
The rapid advancement of Artificial Intelligence Generated Content (AIGC) has produced hyper-realistic synthetic media, raising concerns about authenticity and integrity. Current detection methods are often black-box, lack interpretability, and don't support unified image and video detection, hindering transparency and deployment. To address this, we introduce <strong>IVY-FAKE</strong>, a large-scale, unified dataset for explainable multimodal AIGC detection with over 100,000 annotated training samples and 5,000 evaluation examples, each with natural-language reasoning. We also propose <strong>IVY-XDETECTOR</strong>, a unified vision-language model for state-of-the-art explainable detection of both image and video content.
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="introduction-section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Introduction</h2>
<div class="content has-text-justified">
<p>
Advancements in AIGC, driven by models like DALL-E, Imagen, Stable Diffusion, and SORA, have led to highly realistic synthetic images and videos, posing challenges to content authenticity and public trust. Most current AIGC detection methods are binary classifiers with limited interpretability and often lack support for diverse generators or modalities. While Multimodal Large Language Models (MLLMs) show promise for explainable detection, existing benchmarks are inadequate, often lacking video data or sufficient annotation depth.
</p>
<p>
To overcome these limitations, we introduce <strong>IVY-FAKE</strong>, a comprehensive benchmark with diverse multimodal data and rich, explainable annotations. Building on this, our <strong>IVY-XDETECTOR</strong> model excels at identifying and explaining spatial and temporal generative artifacts in both images and videos. Our key contributions are a unified vision-language detector and the first large-scale benchmark for explainable multimodal AIGC detection.
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="IVY-FAKE Dataset">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">IVY-FAKE Dataset</h2>
<div class="content has-text-justified">
<p>
The <strong>IVY-FAKE dataset</strong> is designed for explainable multimodal AIGC detection, containing 100K+ training images, around 5,000 total test samples. It features diverse content (animals, objects, DeepFakes, etc.) and sources (GANs, Diffusion models, Transformers) and is kept current by collecting new AIGC content.
</p>
<p>
Video data was sourced from public benchmarks like GenVideo and LOKI, and web-crawled platforms, including outputs from models like SORA and Stable Video Diffusion. Image data was similarly collected from public datasets like FakeClue and WildFake, and web sources, covering GANs and Diffusion models. A stratified sampling strategy ensures balanced representation.
</p>
<p>
Explainable annotations were generated using Gemini 2.5 Pro with a knowledge distillation process and a structured template requiring reasoning before conclusion (<code><think>...</think><conclusion>...</conclusion></code>). Gemini was provided with ground-truth labels and a prompt like "This {file_type} is {label}. Explain the reason.". Explanations are categorized into Spatial Features (8 sub-dimensions) and Temporal Features (4 sub-dimensions, video-only). IVY-FAKE uniquely integrates explainable annotations across both image and video modalities, unlike prior datasets which are often limited in scope or modality.
</p>
<div class="has-text-centered" style="margin-top:1em; margin-bottom:1em;">
<img src="static/images/tab_example.png" alt="IVY-FAKE Dataset Overview" style="max-width: 90%; height: auto; border-radius: 12px; box-shadow: 0 4px 10px rgba(0,0,0,0.1);">
<p class="is-size-7"><i>Table 1: Example of IVY-FAKE annotations.</i></p>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section hero is-light" id="detector-section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Detector: IVY-XDETECTOR</h2>
<div class="content has-text-justified">
<p>
This chapter details the IVY-XDETECTOR, a multimodal large language model designed for robust and explainable AIGC detection. The architecture, based on the Qwen2.5-VL paradigm and initialized with Qwen2.5-VL-7B weights. It supports dynamic high-resolution image inputs and processes video frames while preserving temporal information. The model is trained using a progressive two-stage framework: initial video understanding, fine-tuning for AIGC detection, and joint optimization for both detection accuracy and high-quality, human-understandable explanations.
</p>
</div>
<div class="has-text-centered" style="margin-top:1em; margin-bottom:1em;">
<img src="static/images/figure4-Ivy-Fake_00.png" alt="IVY-FAKE Dataset Overview" style="max-width: 90%; height: auto; border-radius: 12px; box-shadow: 0 4px 10px rgba(0,0,0,0.1);">
<p class="is-size-7"><i>The two-stage training pipeline for IVY-XDETECTOR, including general video understanding, detection instruction tuning, and interpretability instruction tuning.</i></p>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="experiments-section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Experiments</h2>
<div class="content has-text-justified">
<p>
This section describes the extensive experiments conducted to evaluate the detection and explanation capabilities of the IVY-XDETECTOR. The model was tested on classification tasks (distinguishing real from fake content) for both images and videos, using metrics like accuracy and F1-score. Reasoning abilities were assessed by comparing model-generated explanations with reference annotations using ROUGE-L scores and an LLM-as-a-judge approach (evaluating completeness, relevance, detail, and explanation quality). The proposed method demonstrated superior performance on various benchmarks for both image and video content classification and produced more transparent reasoning compared to several leading multimodal models.
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section hero is-light" id="conclusion-section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Conclusion</h2>
<div class="content has-text-justified">
<p>
The paper introduces IVY-FAKE, the first unified, large-scale dataset for explainable AI-Generated Content (AIGC) detection across both images and videos, featuring extensive annotated samples with natural-language reasoning. Accompanying this, the Ivy Explainable Detector (IVY-XDETECTOR) is proposed, a vision-language architecture that jointly detects and explains synthetic content. The model sets new state-of-the-art benchmarks in AIGC detection and explainability, and the publicly released resources aim to provide a robust foundation for transparent and trustworthy multimodal analysis.
</p>
</div>
<div class="has-text-centered" style="margin-top:1em; margin-bottom:1em;">
<img src="static/images/figure5-vis_00.png" alt="IVY-FAKE Dataset Overview" style="max-width: 90%; height: auto; border-radius: 12px; box-shadow: 0 4px 10px rgba(0,0,0,0.1);">
<p class="is-size-7"><i>Showcase of the Video artifact detection</i></p>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container">
<div class="columns is-centered">
<div class="column is-full has-text-centered content">
<h2 class="title is-3" id="evaluation">Evaluation Scripts</h2>
<div class="content">
<p class="mt-3">
The official evaluation scripts for IVY-FAKE benchmark can be found at: <br>
<a href="https://github.com/Pi3AI/Ivy-Fake" target="_blank">https://github.com/Pi3AI/Ivy-Fake</a>
</p>
<p>
These scripts support performance evaluation for both image and video modalities, including: <br>
- Automatic metric computation (Acc, F1, ROUGE-L, Bert-Score.) <br>
- GPT-assisted explanation grading <br>
- Final ranking and reporting for MLLM benchmarks
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container">
<div class="columns is-centered">
<div class="column is-full has-text-centered content">
<h2 class="title is-3" id="leaderboard">MLLM Performance Comparison</h2>
<div class="content">
<p class="mt-3">
Performance comparison of models on Image and Video tasks.
“Auto Metrics” include Acc, F1, ROUGE-L, SIM.
“GPT Assisted” includes Comprehensiveness, Relevance, Detail, Explanation.
</p>
<div class="table-container">
<table class="table is-bordered is-striped is-narrow is-hoverable" style="margin-left:auto;margin-right:auto;">
<thead>
<tr>
<th rowspan="3" style="vertical-align: middle;">Model</th>
<th colspan="2" style="text-align:center;">Image</th>
<th colspan="2" style="text-align:center;">Video</th>
<th colspan="2" style="text-align:center;">Overall</th>
</tr>
<tr>
<th style="text-align:center;">Auto Metrics</th>
<th style="text-align:center;">GPT Assisted</th>
<th style="text-align:center;">Auto Metrics</th>
<th style="text-align:center;">GPT Assisted</th>
<th style="text-align:center;">Auto Metrics</th>
<th style="text-align:center;">GPT Assisted</th>
</tr>
<tr>
<th>Acc/F1/ROUGE-L/SIM</th>
<th>Com./Rel./Det./Exp.</th>
<th>Acc/F1/ROUGE-L/SIM</th>
<th>Com./Rel./Det./Exp.</th>
<th>Acc/F1/ROUGE-L/SIM</th>
<th>Com./Rel./Det./Exp.</th>
</tr>
</thead>
<tbody>
<!-- Closed-source -->
<tr><td colspan="7" style="text-align:center;"><i><b>Closed-source MLLMs</b></i></td></tr>
<tr>
<td><b>GPT-4o</b></td>
<td>0.725/0.723/0.108/0.525</td>
<td>2.34/3.20/2.04/3.26</td>
<td>0.448/0.579/0.072/0.451</td>
<td>1.79/2.35/1.67/2.40</td>
<td>0.587/0.663/0.090/0.488</td>
<td>2.07/2.78/1.85/2.83</td>
</tr>
<tr>
<td><b>Gemini-2.5-Flash</b></td>
<td><u>0.747</u>/<u>0.737</u>/<u>0.263</u>/<b>0.733</b></td>
<td><b>3.94</b>/<b>4.11</b>/<b>4.04</b>/<b>4.09</b></td>
<td><u>0.810</u>/<u>0.811</u>/<u>0.246</u>/<u>0.723</u></td>
<td><b>4.00</b>/<b>4.37</b>/<b>4.03</b>/<b>4.36</b></td>
<td><u>0.779</u>/<u>0.776</u>/<u>0.254</u>/<b>0.728</b></td>
<td><b>3.97</b>/<b>4.24</b>/<b>4.04</b>/<b>4.22</b></td>
</tr>
<!-- Open-source -->
<tr><td colspan="7" style="text-align:center;"><i><b>Open-source MLLMs</b></i></td></tr>
<!-- 7B models -->
<tr><td colspan="7"><i>7B-Parameters Models</i></td></tr>
<tr>
<td><b>InternVL3.5-8B</b></td>
<td>0.605/0.602/0.194/0.680</td>
<td>2.83/3.49/2.69/3.32</td>
<td>0.574/0.588/0.188/0.664</td>
<td>2.75/3.35/2.68/3.28</td>
<td>0.589/0.596/0.191/0.672</td>
<td>2.79/3.42/2.69/3.30</td>
</tr>
<tr>
<td><b>MiMo-VL-7B</b></td>
<td>0.662/0.637/0.121/0.593</td>
<td>1.99/2.80/1.85/2.89</td>
<td>0.778/0.783/0.112/0.580</td>
<td>2.04/2.91/1.90/3.19</td>
<td>0.720/0.715/0.116/0.586</td>
<td>2.01/2.86/1.87/3.04</td>
</tr>
<tr>
<td><b>Qwen2.5-VL-7B</b></td>
<td>0.013/0.026/0.006/0.264</td>
<td>1.02/1.03/1.01/1.52</td>
<td>0.092/0.159/0.015/0.280</td>
<td>1.14/1.23/1.12/2.21</td>
<td>0.053/0.096/0.010/0.272</td>
<td>1.08/1.13/1.07/1.86</td>
</tr>
<tr>
<td><b>LLaVA-OneVision-1.5-8B</b></td>
<td>0.500/0.333/0.080/0.499</td>
<td>1.51/2.62/1.49/2.38</td>
<td>0.500/0.333/0.068/0.481</td>
<td>1.49/2.26/1.37/2.19</td>
<td>0.500/0.333/0.074/0.490</td>
<td>1.50/2.44/1.43/2.28</td>
</tr>
<tr>
<td><b>MiniCPM-V-4.5</b></td>
<td>0.666/0.680/0.169/0.637</td>
<td>3.20/3.93/3.06/3.60</td>
<td>0.491/0.505/0.152/0.627</td>
<td>2.83/3.66/2.76/3.36</td>
<td>0.579/0.610/0.161/0.632</td>
<td>3.01/3.80/2.91/3.48</td>
</tr>
<!-- 3B models -->
<tr><td colspan="7"><i>3B-Parameters Models</i></td></tr>
<tr>
<td><b>Qwen2.5-VL-3B</b></td>
<td>0.641/0.612/0.023/0.391</td>
<td>1.19/1.33/1.18/3.28</td>
<td>0.689/0.686/0.017/0.381</td>
<td>1.42/1.56/1.40/3.68</td>
<td>0.665/0.652/0.020/0.386</td>
<td>1.31/1.45/1.29/3.48</td>
</tr>
<tr>
<td><b>Gemma-3-4B-IT</b></td>
<td>0.408/0.477/0.170/0.576</td>
<td>2.55/3.36/2.46/3.11</td>
<td>0.396/0.482/0.149/0.561</td>
<td>2.30/3.03/2.37/2.95</td>
<td>0.402/0.482/0.159/0.568</td>
<td>2.43/3.19/2.42/3.03</td>
</tr>
<tr>
<td><b>InternVL3.5-2B</b></td>
<td>0.602/0.573/0.177/0.648</td>
<td>2.62/3.29/2.51/3.13</td>
<td>0.435/0.459/0.159/0.631</td>
<td>2.46/3.16/2.42/2.99</td>
<td>0.518/0.518/0.168/0.640</td>
<td>2.54/3.22/2.47/3.06</td>
</tr>
<tr>
<td><b>InternVL3.5-4B</b></td>
<td>0.651/0.652/0.190/0.660</td>
<td>3.01/3.68/2.95/3.58</td>
<td>0.614/0.617/0.181/0.653</td>
<td>2.93/3.61/2.83/3.52</td>
<td>0.632/0.635/0.186/0.656</td>
<td>2.97/3.64/2.89/3.55</td>
</tr>
<!-- Baseline (Ivy-xDetector) -->
<tr>
<td><b>Ivy-xDetector</b></td>
<td><b>0.831</b>/<b>0.831</b>/<b>0.283</b>/<u>0.714</u></td>
<td><u>3.54</u>/<u>4.04</u>/<u>3.61</u>/<u>3.85</u></td>
<td><b>0.897</b>/<b>0.897</b>/<b>0.300</b>/<b>0.726</b></td>
<td><u>3.72</u>/<u>4.12</u>/<u>3.75</u>/<u>4.24</u></td>
<td><b>0.864</b>/<b>0.864</b>/<b>0.291</b>/<u>0.720</u></td>
<td><u>3.63</u>/<u>4.08</u>/<u>3.68</u>/<u>4.05</u></td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
</div>
</div>
</section>
<!--BibTex citation -->
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@article{jiang2025ivyfake,
title = {Ivy-Fake: A Unified Explainable Framework and Benchmark for Image and Video AIGC Detection},
author = {Changjiang Jiang and Wenhui Dong and Zhonghao Zhang and Chenyang Si and Fengchang Yu and Wei Peng and Xinbin Yuan and Yifei Bi and Ming Zhao and Zian Zhou and Caifeng Shan},
year = {2025},
url = {https://arxiv.org/abs/2506.00979}
}</code></pre>
</div>
</section>
<!--End BibTex citation -->
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This page was built using the <a href="https://github.com/eliahuhorwitz/Academic-project-page-template" target="_blank">Academic Project Page Template</a> which was adopted from the <a href="https://nerfies.github.io" target="_blank">Nerfies</a> project page.
You are free to borrow the source code of this website, we just ask that you link back to this page in the footer. <br> This website is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
<!-- Statcounter tracking code -->
<!-- You can add a tracker to track page visits by creating an account at statcounter.com -->
<!-- End of Statcounter Code -->
</body>
</html>