eventbench.github.io/index.html at master · XduSyL/eventbench.github.io · GitHub

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
<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="utf-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">

  <meta name="title" content="EventBench">
  <meta name="description" content="">
  <meta name="keywords" content="Event Camera, Event MLLM, Benchmark, Dynamic Vision">
  <meta name="robots" content="index, follow">

  <meta property="og:type" content="article">
  <meta property="og:site_name" content="INSTITUTION_OR_LAB_NAME">
  <meta property="og:title" content="EventBench">
  <meta property="og:description" content="">
  <meta property="og:url" content="https://YOUR_DOMAIN.com">
  <meta property="og:image" content="https://YOUR_DOMAIN.com/static/images/social_preview.png">

  <meta name="twitter:card" content="summary_large_image">
  <meta name="twitter:title" content="EventBench">
  <meta name="twitter:description" content="">
  <meta name="twitter:image" content="https://YOUR_DOMAIN.com/static/images/social_preview.png">

  <title>EventBench</title>

  <link rel="icon" type="image/x-icon" href="static/images/EventBench_logo.png">

  <link href="https://fonts.googleapis.com/css?family=Nunito" rel="stylesheet">

  <link rel="stylesheet" href="static/css/bulma.min.css">
  <link rel="stylesheet" href="static/css/index.css">
  <link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">

  <script defer src="static/js/fontawesome.all.min.js"></script>
  <script defer src="static/js/bulma-carousel.min.js"></script>
  <script defer src="static/js/bulma-slider.min.js"></script>
  <script defer src="static/js/index.js"></script>

  <style>
    body, html,
    h1, h2, h3, h4, h5, h6,
    .title, .subtitle,
    .content, .button,
    .publication-title, .publication-authors,
    .table, .footer, .section,
    .hero, .author-block,
    .work-item, .link-block,
    .more-works-btn, .more-works-dropdown,
    .copy-text, .bibtex-header {
      font-family: 'Nunito', sans-serif !important;
    }
  </style>

</head>

<body>

<button class="scroll-to-top" onclick="scrollToTop()" title="Scroll to top" aria-label="Scroll to top">
  <i class="fas fa-chevron-up"></i>
</button>

<div class="more-works-container">
  <button class="more-works-btn" onclick="toggleMoreWorks()">
    <i class="fas fa-flask"></i>
    More Works
    <i class="fas fa-chevron-down dropdown-arrow"></i>
  </button>

  <div class="more-works-dropdown" id="moreWorksDropdown">
    <div class="dropdown-header">
      <h4>More Works from Our Lab</h4>
      <button class="close-btn" onclick="toggleMoreWorks()">
        <i class="fas fa-times"></i>
      </button>
    </div>
    <div class="works-list">
      <a href="https://openaccess.thecvf.com/content/CVPR2025/papers/Liu_EventGPT_Event_Stream_Understanding_with_Multimodal_Large_Language_Models_CVPR_2025_paper.pdf" class="work-item" target="_blank">
        <div class="work-info">
          <h5>EventGPT: Event Stream Understanding with Multimodal Large Language Models</h5>
          <p>The first event-based multimodal large language model.</p>
          <span class="work-venue">CVPR 2025</span>
        </div>
        <i class="fas fa-external-link-alt"></i>
      </a>
    </div>
  </div>
</div>


<main id="main-content">

<section class="hero" style="padding-bottom:0;">
  <div class="hero-body" style="padding-bottom:0.5rem;">
    <div class="container is-max-desktop">
      <div class="columns is-centered">
        <div class="column has-text-centered">

          <h1 class="title is-1 publication-title" style="line-height:1.25;">

            <div style="display:flex;align-items:center;justify-content:center;gap:10px;white-space:nowrap;">
              <img src="static/images/EventBench_logo.png" style="height:60px;">
              <span>
                <span style="color:rgb(101,114,250);">Event</span>
                <span style="color:rgb(233,99,75);">Bench</span>:
                Towards Comprehensive
              </span>
            </div>

            <div style="white-space:nowrap;margin-top:6px;">
              Benchmarking of Event-based MLLMs
            </div>

          </h1>

          <div class="is-size-5 publication-authors" style="margin-bottom:0.5rem;">
            <span class="author-block">
              <a href="https://scholar.google.com.hk/citations?user=L7OroPIAAAAJ" target="_blank">Shaoyu Liu</a><sup>1,2</sup>,
            </span>
            <span class="author-block">
              <a href="https://scholar.google.com.hk/citations?user=xrYnfwcAAAAJ" target="_blank">Jianing Li</a><sup>2</sup>,
            </span>
            <span class="author-block">
              <a target="_blank">Guanghui Zhao</a><sup>1</sup>,
            </span>
            <span class="author-block">
              <a target="_blank">Yunjian Zhang</a><sup>2</sup>,
            </span>
            <span class="author-block">
              <a href="https://www.au.tsinghua.edu.cn/info/1080/3178.htm" target="_blank">Xiangyang Ji</a><sup>2</sup>
            </span>
          </div>

          <div class="is-size-5 publication-authors" style="margin-top:-0.5rem;margin-bottom:0.5rem;">
            <span class="author-block"><sup>1</sup> Xidian University</span>,
            <span class="author-block"><sup>2</sup> Tsinghua University</span>
            <!-- <span class="eql-cntrb">
              <small><br><sup>*</sup> Corresponding Authors</small>
            </span> -->
          </div>

          <span class="link-block">
            <a href="https://arxiv.org/abs/2511.18448" target="_blank" class="external-link button is-normal is-rounded is-dark">
              <span class="icon">
                <img src="static/images/arxiv.svg" style="height:18px;">
              </span>
              <span>Paper</span>
            </a>
          </span>


          <span class="link-block">
            <a href="https://github.com/XduSyL/EventBench" target="_blank" class="external-link button is-normal is-rounded is-dark">
              <span class="icon"><i class="fab fa-github"></i></span>
              <span>Code</span>
            </a>
          </span>

          <span class="link-block">
            <a href="https://huggingface.co/datasets/XduSyL/EventBench" target="_blank" class="external-link button is-normal is-rounded is-dark">
              <span class="icon"><img src="static/images/hf_logo.svg" style="height:18px;"></span>
              <span>Datasets</span>
            </a>
          </span>

          <span class="link-block">
            <a href="https://huggingface.co/XduSyL/EventGPT-Plus-2B/tree/main" target="_blank" class="external-link button is-normal is-rounded is-dark">
              <span class="icon"><img src="static/images/hf_logo.svg" style="height:18px;"></span>
              <span>Model</span>
            </a>
          </span>

        </div>
      </div>
    </div>
  </div>
</section>


<div class="container is-max-desktop" style="padding:0;margin-top:0.25rem;margin-bottom:1rem;">
  <img src="static/images/overview.png" style="width:100%;display:block;">

  <h2 class="subtitle"
      style="margin-top:1.2rem;color:#000;text-align:justify;font-size:1.05rem;line-height:1.55;">

      The comprehensive <strong>EventBench</strong> covers 8 diverse task metrics for systematically evaluating the capabilities of <strong>event-based MLLMs</strong>.
      These metrics can be broadly categorized into three types:

      <span style="color:rgb(101,114,250); font-weight:bold;">
        understanding
      </span>
      (detailed understanding, casual reasoning),

      <span style="color:rgb(233,99,75); font-weight:bold;">
        recognition
      </span>
      (action recognition, gesture recognition, and event OCR),

      and
      <span style="color:rgb(198,95,16); font-weight:bold;">
        spatial reasoning
      </span>
      (spatial relationship, absolute distance, and object counting).
  </h2>
</div>


<section class="section hero is-light">
  <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">Abstract</h2>
        <div class="content has-text-justified">
          <p style="text-align: justify; line-height: 1.55; font-size: 1.05rem;">

            Multimodal large language models (MLLMs) have made significant advancements in event-based vision.
            However, the comprehensive evaluation of these models' capabilities within a unified benchmark remains a crucial yet largely unexplored aspect.
            In this paper, we introduce <strong>EventBench</strong>, which presents an evaluation benchmark covering 8 diverse task metrics and a large-scale event stream dataset.
            Our work distinguishes existing event-based benchmarks in four key aspects:
            <strong>i) Openness in accessibility</strong>, releasing all raw event streams and task instructions across 8 evaluation metrics;
            <strong>ii) Diversity in task coverage</strong>, covering diverse understanding, recognition, and spatial reasoning tasks, enabling comprehensive evaluation of model capability;
            <strong>iii) Integration in spatial dimensions</strong>, pioneering the design of 3D spatial reasoning task for event-based MLLMs;
            <strong>iv) Scale in data volume</strong>, with an accompanying training set of over one million event–text pairs supporting large-scale training and evaluation.
            With EventBench, we evaluate state-of-the-art closed-source models such as GPT-5 and Gemini-2.5 Pro, leading open-source models including Qwen2.5-VL and InternVL3, as well as event-based MLLMs like EventGPT that directly process raw event inputs.
            Extensive evaluation reveals that current event-based MLLMs perform well in event stream understanding, yet they still struggle with fine-grained recognition and spatial reasoning.

          </p>
        </div>
      </div>
    </div>
  </div>
</section>

<!-- ===================================== -->
<!-- Radar Chart + Main Results Table -->
<!-- ===================================== -->
<!-- ===================================== -->
<!-- Radar Figure + Main Results Table -->
<!-- ===================================== -->
<section class="section">
  <div class="container is-max-desktop">

    <h2 class="title is-3 has-text-centered">Main Results on EventBench</h2>

    <!-- Radar Figure -->
    <div style="text-align:center; margin-bottom: 1.5rem;">
      <img src="static/images/lidar_graph.png"
           alt="Radar Performance Graph"
           style="width:50%; height:auto; box-shadow:none; background:none; border-radius:0;">
    </div>

    <div class="table-container">
      <table class="table is-bordered is-striped is-hoverable"
             style="width:100%; font-size:0.9rem;">

        <!-- Header -->
        <thead>
          <tr>
            <th rowspan="2" style="vertical-align: middle; text-align:center;">Model</th>
            <th rowspan="2" style="vertical-align: middle; text-align:center;">Params</th>
            <th rowspan="2" style="vertical-align: middle; text-align:center;">Backbone</th>

            <th colspan="2" style="text-align:center;">Understanding</th>
            <th colspan="3" style="text-align:center;">Recognition</th>
            <th colspan="3" style="text-align:center;">Spatial Reasoning</th>

            <th rowspan="2" style="vertical-align: middle; text-align:center;">Overall</th>
          </tr>

          <tr>
            <th>DU</th><th>CR</th>
            <th>AR</th><th>GR</th><th>E-OCR</th>
            <th>SR</th><th>AD</th><th>OC</th>
          </tr>
        </thead>

        <tbody>

          <!-- Close-Source -->
          <tr style="background:#f2f2f2;">
            <td colspan="12"><strong>● Close-Source MLLMs</strong></td>
          </tr>

          <tr>
            <td>GPT-5</td><td>-</td><td>-</td>
            <td>68.5</td><td><strong>70.1</strong></td>
            <td>46.0</td><td>49.8</td><td>22.2</td>
            <td><strong>47.8</strong></td><td><strong>30.7</strong></td><td><strong>25.6</strong></td>
            <td><strong>45.1</strong></td>
          </tr>

          <tr>
            <td>GPT-4o</td><td>-</td><td>-</td>
            <td>62.5</td><td>65.8</td>
            <td><strong>55.3</strong></td><td>48.6</td><td>19.1</td>
            <td>45.5</td><td>17.8</td><td>21.5</td>
            <td>42.0</td>
          </tr>

          <tr>
            <td>Gemini-2.5 Pro</td><td>-</td><td>-</td>
            <td>61.3</td><td>63.9</td>
            <td>44.1</td><td><strong>49.9</strong></td><td>19.1</td>
            <td>40.5</td><td>17.5</td><td>16.9</td>
            <td>39.2</td>
          </tr>

          <tr>
            <td>Gemini-2.5 Flash</td><td>-</td><td>-</td>
            <td>54.7</td><td>61.1</td>
            <td>42.7</td><td>43.4</td><td>21.0</td>
            <td>25.6</td><td>15.6</td><td>14.5</td>
            <td>34.8</td>
          </tr>

          <tr>
            <td>Doubao-Seed-1.6 Vision</td><td>-</td><td>-</td>
            <td><strong>70.3</strong></td><td>65.2</td>
            <td>46.6</td><td>45.4</td><td><strong>45.1</strong></td>
            <td>11.6</td><td>36.3</td><td>23.0</td>
            <td>42.9</td>
          </tr>

          <!-- Open-Source SFT -->
          <tr style="background:#f2f2f2;">
            <td colspan="12"><strong>● Open-Source MLLMs (SFT)</strong></td>
          </tr>

          <tr>
            <td>Qwen2.5-VL</td><td>3B</td><td>Qwen2.5</td>
            <td>59.8</td><td>65.6</td>
            <td>52.9</td><td>35.2</td><td><strong>45.6</strong></td>
            <td><strong>44.9</strong></td><td>31.9</td><td>59.9</td>
            <td>49.5</td>
          </tr>

          <tr>
            <td>InternVL3.5</td><td>4B</td><td>Qwen3</td>
            <td>73.1</td><td><strong>75.0</strong></td>
            <td><strong>69.3</strong></td><td><strong>45.2</strong></td><td>38.9</td>
            <td>40.5</td><td><strong>46.8</strong></td><td><strong>61.1</strong></td>
            <td><strong>56.2</strong></td>
          </tr>

          <tr>
            <td>MiniCPM-V-4</td><td>4.1B</td><td>MiniCPM4</td>
            <td><strong>72.9</strong></td><td>63.9</td>
            <td>58.9</td><td>37.1</td><td>44.4</td>
            <td>44.2</td><td>43.1</td><td>55.6</td>
            <td>52.5</td>
          </tr>

          <tr>
            <td>MiMo-VL</td><td>7B</td><td>MiMo</td>
            <td>71.8</td><td>68.3</td>
            <td>57.3</td><td>38.6</td><td>31.5</td>
            <td>44.5</td><td>45.2</td><td>61.0</td>
            <td>52.3</td>
          </tr>

          <!-- Event-Based Models -->
          <tr style="background:#f2f2f2;">
            <td colspan="12"><strong>● Open-Source Event-Based MLLMs</strong></td>
          </tr>

          <tr>
            <td>EventGPT</td><td>7B</td><td>Vicuna-1.5</td>
            <td>76.8</td><td>79.2</td>
            <td>78.6</td><td>57.2</td><td>52.4</td>
            <td>49.8</td><td>52.6</td><td>61.5</td>
            <td>63.5</td>
          </tr>

          <tr>
            <td><strong>EventGPT+(B)</strong></td><td>2B</td><td>Qwen2.5</td>
            <td>78.3</td><td>78.2</td>
            <td>79.8</td><td>55.5</td><td>53.6</td>
            <td>50.2</td><td>51.9</td><td>62.8</td>
            <td>63.8</td>
          </tr>

          <tr>
            <td><strong>EventGPT+(L)</strong></td><td>7B</td><td>Qwen2.5</td>
            <td><strong>79.1</strong></td><td><strong>79.6</strong></td>
            <td><strong>81.6</strong></td><td><strong>58.7</strong></td><td><strong>55.4</strong></td>
            <td><strong>53.2</strong></td><td><strong>52.2</strong></td><td><strong>64.6</strong></td>
            <td><strong>65.5</strong></td>
          </tr>

        </tbody>
      </table>
    </div>

  </div>
</section>


<!-- =============================== -->
<!-- Dataset Statistics Section -->
<!-- =============================== -->
<!-- =============================== -->
<!-- Dataset Statistics Section -->
<!-- =============================== -->
<!-- =============================== -->
<!-- Dataset Statistics Section -->
<!-- =============================== -->
<section class="section" style="padding-top: 2rem; padding-bottom: 2rem;">

  <!-- Medium-wide container -->
  <div style="max-width: 1000px; margin: 0 auto;">

    <h2 class="title is-3 has-text-centered" style="margin-bottom: 1.2rem;">
      Dataset Statistics
    </h2>

    <div class="has-text-centered">
      <img src="static/images/data_statistics.png"
           alt="Dataset Statistics"
           style="
             width: 100%;
             max-width: 1150px;
             display: block;
             margin-left: auto;
             margin-right: auto;
             border-radius: 6px;
           ">
    </div>

    <!-- Caption -->
    <p style="
        margin-top: 1rem;
        text-align: justify;
        font-size: 1.05rem;
        line-height: 1.55;
        color: #000;">
        Data statistics of our EventBench. (a) Task and category distribution across three groups: understanding (i.e., DU and CR), recognition (i.e., AR, GR, and E-OCR), and spatial reasoning (i.e., SR, AD, and OC). (b) Sample statistics for each task category. (c) Comparison with existing event-based benchmarks across multiple dimensions (i.e., modality, metric type, data source, size, and temporal span). Note that EventBench provides a comprehensive benchmark for systematically evaluating the capabilities of event-based MLLMs.
    </p>

  </div>
</section>


<section class="section" id="BibTeX">
  <div class="container is-max-desktop content">
    <div class="bibtex-header">
      <h2 class="title">BibTeX</h2>
      <button class="copy-bibtex-btn" onclick="copyBibTeX()">
        <i class="fas fa-copy"></i>
        <span class="copy-text">Copy</span>
      </button>
    </div>

    <pre id="bibtex-code"><code>@article{liu2025eventbench,
  title={EventBench: Towards Comprehensive Benchmarking of Event-based MLLMs},
  author={Liu, Shaoyu and Li, Jianing and Zhao, Guanghui and Zhang, Yunjian and Ji, Xiangyang},
  journal={arXiv preprint arXiv:2511.18448},
  year={2025}
}</code></pre>
  </div>
</section>


<footer class="footer">
  <div class="container">
    <div class="columns is-centered">
      <div class="column is-8">
        <div class="content">
          <p>
            This page is based on the Academic Project Page Template and Nerfies.
          </p>
        </div>
      </div>
    </div>
  </div>
</footer>

</main>

</body>
</html>