Skip to content

Commit e453844

Browse files
ZhihaoPENG-CityUZhihaoPENG-CityU
authored
Update index.html
1 parent 4b622e9 commit e453844

1 file changed

Lines changed: 82 additions & 88 deletions

File tree

‎index.html‎

Lines changed: 82 additions & 88 deletions
Original file line numberDiff line numberDiff line change
@@ -285,7 +285,7 @@ <h2 class="title is-4">Performance of different MLLMs on open-ended VQA of OmniB
285285
<h2 class="title is-4">Diverse Modality Evaluation.</h2>
286286
<div class="content has-text-centered">
287287
<img src="static/images/OmniBrainBench_analysis_DiffModality.png" alt="algebraic reasoning" width="75%" class="center">
288-
<p>Table 3: Diverse Modality Evaluation.</p>
288+
<p>Figure 1: Diverse Modality Evaluation.</p>
289289
</div>
290290
</div>
291291
</div>
@@ -296,7 +296,7 @@ <h2 class="title is-4">Diverse Modality Evaluation.</h2>
296296
<h2 class="title is-4">Performance of models on different numbers of images.</h2>
297297
<div class="content has-text-centered">
298298
<img src="static/images/OmniBrainBench_analysis_DiffImages.png" alt="algebraic reasoning" width="80%" class="center">
299-
<p>Figure 1: Performance of models on different numbers of images.</p>
299+
<p>Figure 2: Performance of models on different numbers of images.</p>
300300
</div>
301301
</div>
302302
</div>
@@ -420,92 +420,86 @@ <h2 class="title is-3">Case Study</h2>
420420
These case studies emphasize the need for improved medical knowledge integration and enhanced perceptual capabilities to bridge the gap between current MLLM performance and clinical requirements.
421421
</p> -->
422422
</div>
423-
<!-- <div class="columns is-multiline">
424-
<div class="column is-half">
425-
<figure class="image">
426-
<img src="static/images/Case Study_01.jpg" alt="Case Study 01" class="center">
427-
<figcaption>Figure 7: Correct sample from Gemini-2.5-Pro.</figcaption>
428-
</figure>
429-
</div>
430-
<div class="column is-half">
431-
<figure class="image">
432-
<img src="static/images/Case Study_02.jpg" alt="Case Study 02" class="center">
433-
<figcaption>Figure 8: Correct sample from Gemini-2.5-Pro.</figcaption>
434-
</figure>
435-
</div>
436-
<div class="column is-half">
437-
<figure class="image">
438-
<img src="static/images/Case Study_03.jpg" alt="Case Study 03" class="center">
439-
<figcaption>Figure 9: Correct sample from Gemini-2.5-Pro.</figcaption>
440-
</figure>
441-
</div>
442-
<div class="column is-half">
443-
<figure class="image">
444-
<img src="static/images/Case Study_04.jpg" alt="Case Study 04" class="center">
445-
<figcaption>Figure 10: Correct sample from GPT-4o.</figcaption>
446-
</figure>
447-
</div>
448-
<div class="column is-half">
449-
<figure class="image">
450-
<img src="static/images/Case Study_05.jpg" alt="Case Study 05" class="center">
451-
<figcaption>Figure 11: Correct sample from GPT-4o.</figcaption>
452-
</figure>
453-
</div>
454-
<div class="column is-half">
455-
<figure class="image">
456-
<img src="static/images/Case Study_06.jpg" alt="Case Study 06" class="center">
457-
<figcaption>Figure 12: Correct sample from GPT-4o.</figcaption>
458-
</figure>
459-
</div>
460-
<div class="column is-half">
461-
<figure class="image">
462-
<img src="static/images/Case Study_07.jpg" alt="Case Study 07" class="center">
463-
<figcaption>Figure 13: Error sample demonstrating Perceptual Errors (QvQ-72B).</figcaption>
464-
</figure>
465-
</div>
466-
<div class="column is-half">
467-
<figure class="image">
468-
<img src="static/images/Case Study_08.jpg" alt="Case Study 08" class="center">
469-
<figcaption>Figure 14: Error sample demonstrating Perceptual Errors (HuatuoGPT-Vision-34B).</figcaption>
470-
</figure>
471-
</div>
472-
<div class="column is-half">
473-
<figure class="image">
474-
<img src="static/images/Case Study_09.jpg" alt="Case Study 09" class="center">
475-
<figcaption>Figure 15: Error sample demonstrating Lack of Knowledge (QvQ-72B).</figcaption>
476-
</figure>
477-
</div>
478-
<div class="column is-half">
479-
<figure class="image">
480-
<img src="static/images/Case Study_10.jpg" alt="Case Study 10" class="center">
481-
<figcaption>Figure 16: Error sample demonstrating Lack of Knowledge (HuatuoGPT-Vision-34B).</figcaption>
482-
</figure>
483-
</div>
484-
<div class="column is-half">
485-
<figure class="image">
486-
<img src="static/images/Case Study_11.jpg" alt="Case Study 11" class="center">
487-
<figcaption>Figure 17: Error sample demonstrating Irrelevant Response (LLaVA-Med).</figcaption>
488-
</figure>
489-
</div>
490-
<div class="column is-half">
491-
<figure class="image">
492-
<img src="static/images/Case Study_12.jpg" alt="Case Study 12" class="center">
493-
<figcaption>Figure 18: Error sample demonstrating Irrelevant Response (ColonGPT).</figcaption>
494-
</figure>
495-
</div>
496-
<div class="column is-half">
497-
<figure class="image">
498-
<img src="static/images/Case Study_13.jpg" alt="Case Study 13" class="center">
499-
<figcaption>Figure 19: Error sample demonstrating Refusal to Answer (GPT-4o).</figcaption>
500-
</figure>
501-
</div>
502-
<div class="column is-half">
503-
<figure class="image">
504-
<img src="static/images/Case Study_14.jpg" alt="Case Study 14" class="center">
505-
<figcaption>Figure 20: Error sample demonstrating Refusal to Answer (Grok-3).</figcaption>
506-
</figure>
507-
</div>
508-
</div> -->
423+
<div class="columns is-multiline">
424+
<!-- GPT-5 -->
425+
<div class="column is-half">
426+
<figure class="image">
427+
<img src="static/images/GPT-5 closed-ended VQA.png" alt="GPT-5 Closed-ended VQA Samples" class="center">
428+
<figcaption>Figure 3: Correct/Error samples in GPT-5 closed-ended VQA.</figcaption>
429+
</figure>
430+
</div>
431+
<div class="column is-half">
432+
<figure class="image">
433+
<img src="static/images/GPT-5 open-ended VQA.png" alt="GPT-5 Open-ended VQA Samples" class="center">
434+
<figcaption>Figure 4: Correct/Error samples in GPT-5 open-ended VQA.</figcaption>
435+
</figure>
436+
</div>
437+
<!-- Claude-4.5-Sonnet -->
438+
<div class="column is-half">
439+
<figure class="image">
440+
<img src="static/images/Claude-4.5-Sonnet closed-ended VQA.png" alt="Claude-4.5-Sonnet Closed-ended VQA Samples" class="center">
441+
<figcaption>Figure 5: Correct/Error samples in Claude-4.5-Sonnet closed-ended VQA.</figcaption>
442+
</figure>
443+
</div>
444+
<div class="column is-half">
445+
<figure class="image">
446+
<img src="static/images/Claude-4.5-Sonnet open-ended VQA.png" alt="Claude-4.5-Sonnet Open-ended VQA Samples" class="center">
447+
<figcaption>Figure 6: Correct/Error samples in Claude-4.5-Sonnet open-ended VQA.</figcaption>
448+
</figure>
449+
</div>
450+
<!-- Gemini-2.5-Pro -->
451+
<div class="column is-half">
452+
<figure class="image">
453+
<img src="static/images/Gemini-2.5-Pro closed-ended VQA.png" alt="Gemini-2.5-Pro Closed-ended VQA Samples" class="center">
454+
<figcaption>Figure 7: Correct/Error samples in Gemini-2.5-Pro closed-ended VQA.</figcaption>
455+
</figure>
456+
</div>
457+
<div class="column is-half">
458+
<figure class="image">
459+
<img src="static/images/Gemini-2.5-Pro open-ended VQA.png" alt="Gemini-2.5-Pro Open-ended VQA Samples" class="center">
460+
<figcaption>Figure 8: Correct/Error samples in Gemini-2.5-Pro open-ended VQA.</figcaption>
461+
</figure>
462+
</div>
463+
<!-- Deepseek-V3.1 -->
464+
<div class="column is-half">
465+
<figure class="image">
466+
<img src="static/images/Deepseek-V3.1 closed-ended VQA.png" alt="Deepseek-V3.1 Closed-ended VQA Samples" class="center">
467+
<figcaption>Figure 9: Correct/Error samples in Deepseek-V3.1 closed-ended VQA.</figcaption>
468+
</figure>
469+
</div>
470+
<div class="column is-half">
471+
<figure class="image">
472+
<img src="static/images/Deepseek-V3.1 open-ended VQA.png" alt="Deepseek-V3.1 Open-ended VQA Samples" class="center">
473+
<figcaption>Figure 10: Correct/Error samples in Deepseek-V3.1 open-ended VQA.</figcaption>
474+
</figure>
475+
</div>
476+
<!-- Qwen3-VL-30B -->
477+
<div class="column is-half">
478+
<figure class="image">
479+
<img src="static/images/Qwen3-VL-30B closed-ended VQA.png" alt="Qwen3-VL-30B Closed-ended VQA Samples" class="center">
480+
<figcaption>Figure 11: Correct/Error samples in Qwen3-VL-30B closed-ended VQA.</figcaption>
481+
</figure>
482+
</div>
483+
<div class="column is-half">
484+
<figure class="image">
485+
<img src="static/images/Qwen3-VL-30B open-ended VQA.png" alt="Qwen3-VL-30B Open-ended VQA Samples" class="center">
486+
<figcaption>Figure 12: Correct/Error samples in Qwen3-VL-30B open-ended VQA.</figcaption>
487+
</figure>
488+
</div>
489+
<!-- Lingshu-32B -->
490+
<div class="column is-half">
491+
<figure class="image">
492+
<img src="static/images/Lingshu-32B closed-ended VQA.png" alt="Lingshu-32B Closed-ended VQA Samples" class="center">
493+
<figcaption>Figure 13: Correct/Error samples in Lingshu-32B closed-ended VQA.</figcaption>
494+
</figure>
495+
</div>
496+
<div class="column is-half">
497+
<figure class="image">
498+
<img src="static/images/Lingshu-32B open-ended VQA.png" alt="Lingshu-32B Open-ended VQA Samples" class="center">
499+
<figcaption>Figure 14: Correct/Error samples in Lingshu-32B open-ended VQA.</figcaption>
500+
</figure>
501+
</div>
502+
</div>
509503
</div>
510504
</div>
511505
</div>

0 commit comments

Comments
 (0)