-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
2763 lines (2658 loc) · 133 KB
/
index.html
File metadata and controls
2763 lines (2658 loc) · 133 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
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html>
<head>
<title>Goji-Benchmark</title>
<style>
.hidden {
display: none;
}
</style>
<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
<script src="https://kit.fontawesome.com/f8ddf9854a.js" crossorigin="anonymous"></script>
<meta charset="utf-8">
<meta name="description" content="Ophthalmology Benchmarks on Large Language Models">
<meta name="keywords"
content="Goji, Vision Language Model, Large Language Model, Large Multimodal Model, artificial intelligence, AI, AGI, artificial general intelligence">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title> Goji: Ophthalmology Benchmarks on Large Multimodal Models</title>
<link rel="icon" href="./assets/webpage/static/images/goji_logo.png">
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro" rel="stylesheet">
<link rel="stylesheet" href="./assets/webpage/static/css/bulma.min.css">
<link rel="stylesheet" href="./assets/webpage/static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./assets/webpage/static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./assets/webpage/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="./assets/webpage/static/css/index.css">
<link rel="stylesheet" href="./assets/webpage/static/css/leaderboard.css">
<style>
.no-sort {
pointer-events: none !important;
cursor: default !important;
}
.no-sort:hover {
background-color: inherit !important;
}
</style>
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./assets/webpage/static/js/fontawesome.all.min.js"></script>
<script src="./assets/webpage/static/js/bulma-carousel.min.js"></script>
<script src="./assets/webpage/static/js/bulma-slider.min.js"></script>
<script src="./assets/webpage/static/js/index.js"></script>
<script src="./assets/webpage/static/js/question_card.js"></script>
<!-- Data files - uncomment when available -->
<!-- <script src="./data/results/data_setting.js" defer></script> -->
<!-- <script src="./data/results/model_scores.js" defer></script> -->
<!-- <script src="./visualizer/data/data_public.js" defer></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">
<h1 class="title is-1 publication-title is-bold">
<img src="./assets/webpage/static/images/goji_logo.png" style="width:1em;vertical-align: middle"
alt="Logo" />
<span class="lmod" style="vertical-align: middle">LMOD+</span>
</h1>
<h2 class="subtitle is-3 publication-subtitle">
A Comprehensive Multimodal Dataset and Benchmark for Developing and Evaluating Multimodal Large Language
Models in Ophthalmology
</h2>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://kfzyqin.github.io/" style="text-decoration: none; color: inherit;">Zhenyue
Qin<sup style="color:#6fbf73;">†,1</sup></a>,
</span>
<span class="author-block">
Yang Liu<sup style="color:#87A2FF;">†,2</sup>,
</span>
<span class="author-block">
Yu Yin<sup style="color:#ffac33;">3</sup>,
</span>
<span class="author-block">
Jinyu Ding<sup style="color:#6fbf73;">1</sup>,
</span>
<span class="author-block">
Casper Zhang<sup style="color:#6fbf73;">1</sup>,
</span><br>
<span class="author-block">Anran Li<sup style="color:#6fbf73;">1</sup>,</span>
<span class="author-block">
Dylan Campbell<sup style="color:#87A2FF;">2</sup>,
</span>
<span class="author-block">
Xuansheng Wu<sup style="color:#ed4b82;">4</sup>,
</span>
<span class="author-block">
Ke Zou<sup style="color:#007bff;">5</sup>,
</span>
<span class="author-block">Tiarnan D. L. Keenan<sup style="color:#28a745;">6</sup>,</span><br>
<span class="author-block">Emily Y. Chew<sup style="color:#28a745;">6</sup>,</span>
<span class="author-block">Zhiyong Lu<sup style="color:#28a745;">6</sup>,</span>
<span class="author-block">Yih-Chung Tham<sup style="color:#007bff;">5</sup>,</span>
<span class="author-block">Ninghao Liu<sup style="color:#ed4b82;">4</sup>,</span>
<span class="author-block">Xiuzhen Zhang<sup style="color:#dc3545;">7</sup>,</span>
<span class="author-block">
<a href="https://medicine.yale.edu/profile/qingyu-chen/" style="text-decoration: none; color: inherit;">
Qingyu Chen<sup style="color:#6fbf73;">*,1</sup> </a>
</span>
<br>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup style="color:#6fbf73;">1</sup>Yale University,</span>
<span class="author-block"><sup style="color:#87A2FF;">2</sup>Australian National University,</span>
<span class="author-block"><sup style="color:#ffac33;">3</sup>Imperial College London,</span>
<span class="author-block"><sup style="color:#ed4b82;">4</sup>University of Georgia,</span>
<span class="author-block"><sup style="color:#007bff;">5</sup>National University of
Singapore,</span><br>
<span class="author-block"><sup style="color:#28a745;">6</sup>National Institutes of Health,</span>
<span class="author-block"><sup style="color:#dc3545;">7</sup>RMIT University</span>
</div>
<br>
<div class="is-size-5 publication-authors">
<span class="author-block">†Equal contribution</span><br>
<span class="author-block">*Corresponding to:</span>
<span class="author-block"><a href="mailto:qingyu.chen@yale.edu">qingyu.chen@yale.edu</a></span>
</div>
<br>
<div class="publication-links">
<a href="https://arxiv.org/pdf/2410.01620" target="_blank" class="button-link">
<i class="fas fa-file-pdf"></i> Paper
</a>
<!-- <a href="#" target="_blank" class="button-link">
<i class="fab fa-github"></i> Code
</a> -->
<a href="https://yaleedu-my.sharepoint.com/:f:/r/personal/zhenyue_qin_yale_edu/Documents/Research/Data/LMOD/Processed-Data?csf=1&web=1&e=jf7rso"
target="_blank" class="button-link">
<i class="fas fa-database"></i> Dataset
</a>
<!-- <a href="#leaderboard" class="button-link">
<i class="fas fa-trophy"></i> Leaderboard
</a> -->
<a href="https://www.linkedin.com/feed/update/urn:li:activity:7322730725097582611/" target="_blank"
class="button-link">
<i class="fas fa-newspaper"></i> News
</a>
</div>
</div>
</div>
</div>
</div>
</section>
<style>
.center {
display: block;
margin-left: auto;
margin-right: auto;
width: 80%;
}
.publication-links {
display: flex;
justify-content: center;
align-items: center;
gap: 12px;
flex-wrap: wrap;
margin-top: 20px;
}
.button-link {
display: inline-flex;
align-items: center;
gap: 8px;
padding: 10px 20px;
background-color: #2c3e50;
color: white !important;
text-decoration: none !important;
border-radius: 25px;
font-size: 14px;
font-weight: 500;
transition: all 0.3s ease;
box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
}
.button-link:hover {
background-color: #34495e;
transform: translateY(-2px);
box-shadow: 0 4px 8px rgba(0, 0, 0, 0.2);
color: white !important;
}
.button-link i {
font-size: 16px;
}
/* Related Work Table Styles */
.table-container {
box-shadow: 0 2px 8px rgba(0, 0, 0, 0.1);
border-radius: 8px;
overflow: hidden;
margin: 20px 0;
overflow-x: auto;
}
.table th {
background-color: #f8f9fa !important;
font-weight: 700;
color: #2c3e50;
border: 1px solid #dee2e6;
padding: 8px 4px;
font-size: 13px;
white-space: nowrap;
}
.table td {
padding: 8px 4px;
font-size: 12px;
vertical-align: middle;
text-align: center;
border: 1px solid #dee2e6;
}
.table tbody tr:hover {
background-color: #f1f3f4;
}
.table td:first-child {
text-align: left;
font-weight: 500;
min-width: 120px;
}
/* Checkmark and X styling */
.table td:contains("✓") {
color: #28a745;
font-weight: bold;
}
.table td:contains("✗") {
color: #dc3545;
font-weight: bold;
}
/* Leaderboard Table Styles */
#table1 {
width: 100%;
border-collapse: collapse;
margin: 20px 0;
font-size: 14px;
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
box-shadow: 0 4px 12px rgba(0, 0, 0, 0.15);
border-radius: 8px;
overflow: hidden;
}
#table1 thead th {
text-align: center;
font-weight: 700;
border: 1px solid #dee2e6;
position: relative;
}
#table1 tbody tr {
transition: all 0.2s ease;
border-bottom: 1px solid #dee2e6;
}
#table1 tbody tr:hover {
background-color: #f8f9fa !important;
transform: translateY(-1px);
box-shadow: 0 2px 8px rgba(0, 0, 0, 0.1);
}
#table1 tbody td {
border: 1px solid #e9ecef;
transition: all 0.2s ease;
}
#table1 tbody td:first-child {
text-align: left;
font-weight: 600;
min-width: 200px;
}
/* Performance highlighting */
.best-performance {
background: linear-gradient(135deg, #fff3cd 0%, #ffeaa7 100%);
font-weight: 700;
color: #856404;
border: 2px solid #f39c12 !important;
}
.second-performance {
background: linear-gradient(135deg, #e8f4fd 0%, #b3d9ff 100%);
font-weight: 600;
color: #2c5aa0;
text-decoration: underline;
}
</style>
<section class="section" style="padding-top: 2rem;">
<div class="container">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<!-- <h2 class="title is-3">🔔News</h2>
<div class="content has-text-justified">
<p>
<b>🚀[2024-01-31]: We added Human Expert performance on the <a href="#leaderboard">Leaderboard</a>!🌟</b>
</p>
<p>
<b>🔥[2023-12-04]: Our evaluation server for the test set is now available on <a
href="https://eval.ai/web/challenges/challenge-page/2179/overview"><b>EvalAI</b></a>. We welcome all
submissions and look forward to your participation! 😆</b>
</p>
</div> -->
<div class="content has-text-justified">
<p>
Vision-threatening eye diseases pose a major global health burden, affecting more than 2.2 billion people
worldwide.
In the United States alone, over 90 million people are at high risk for vision loss, yet many remain
undiagnosed
or are diagnosed too late for effective treatment. Up to 50% of patients with diabetic retinopathy do not
receive
timely eye examinations, highlighting critical gaps in screening and management.
</p>
<p>
While artificial intelligence offers promising solutions through multimodal large language models (MLLMs),
a major challenge is the lack of unified, comprehensive benchmarks for ophthalmology. Most existing
benchmarks
were designed for earlier CNN-based models or focus on text-only tasks, failing to reflect real-world
ophthalmic
practice where medical imaging is indispensable.
</p>
<p>
This work presents <b>LMOD+</b>, a significantly enhanced version of our large-scale multimodal
ophthalmology benchmark,
comprising <b>32,633 images</b> with multi-granular annotations across <b>12 common ophthalmic
conditions</b> and
<b>5 imaging modalities</b>. Our key contributions include:
</p>
<ul>
<li><b>Comprehensive Dataset:</b> Our dataset encompasses 32,633 high-quality images, featuring an
extensive
collection of color fundus photographs that covers diverse pathological conditions</li>
<li><b>Diverse Tasks:</b> Comprehensive evaluation across 12 binary eye condition diagnosis tasks,
multi-class disease diagnosis, severity classification, and demographic prediction to assess potential
bias</li>
<li><b>Extensive Evaluation:</b> Systematic assessment of 24 state-of-the-art MLLMs,
including recent advanced models such as InternVL, Qwen, and DeepSeek series</li>
<li><b>Public Resources:</b> Full dataset release with dynamic leaderboard and evaluation pipeline to
support ongoing benchmarking and model development</li>
</ul>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<!-- DATASET SECTION -->
<!-- <section class="hero is-light is-small">
<div class="hero-body has-text-centered">
<h1 class="title is-1 lmod">
<img src="./assets/webpage/static/images/goji_logo.png" style="width:1em;vertical-align: middle" alt="Logo" />
<span class="lmod" style="vertical-align: middle">LMOD+</span>
</h1>
</div>
</section> -->
<!-- LMOD+ DATASET SECTION -->
<section class="hero is-primary is-small" style="background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);">
<div class="hero-body has-text-centered">
<h1 class="title is-1 lmod">
<img src="./assets/webpage/static/images/goji_logo.png" style="width:1em;vertical-align: middle" alt="Logo" />
<span class="lmod" style="vertical-align: middle">LMOD+ Dataset</span>
</h1>
</div>
</section>
<section class="section">
<div class="container">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<div class="content has-text-justified">
<p>
Comparison of existing general-domain and ophthalmology-specific benchmarks for evaluating large
vision-language models, highlighting their supported modalities, coverage of image types,
and evaluation perspectives.
</p>
</div>
<div class="table-container">
<table id="related-work-table"
class="table is-bordered is-striped is-narrow is-hoverable is-fullwidth js-sort-none">
<thead>
<tr style="background-color: #f5f5f5;">
<th rowspan="2" class="has-text-centered js-sort-none no-sort" style="vertical-align: middle;">
<strong>Benchmarks</strong>
</th>
<th colspan="2" class="has-text-centered js-sort-none"><strong>Modalities</strong></th>
<th colspan="5" class="has-text-centered js-sort-none"><strong>Image Types</strong></th>
<th colspan="2" class="has-text-centered js-sort-none"><strong>Evaluation Perspectives</strong></th>
</tr>
<tr style="background-color: #f5f5f5;">
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Images</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Texts</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Surgical Scenes
(SS)</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Optical Coherence
Tomography
(OCT)</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Scanning Laser
Ophthalmoscopy
(SLO)</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Lens Photographs
(LP)</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Color Fundus Photographs
(CFP)</strong>
</th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Anatomical
Understanding</strong></th>
<th class="has-text-centered js-sort-none" style="font-size: 12px;"><strong>Diagnosis
Analysis</strong></th>
</tr>
</thead>
<tbody>
<tr style="background-color: #f8f9fa; font-weight: 600;">
<td colspan="10" class="has-text-centered" style="background-color: #e9ecef;"><strong>General-Domain
Benchmarks</strong></td>
</tr>
<tr>
<td><strong>MMMU</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>MME-RealWorld</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>UNK-VQA</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>MMBench</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>MathVista</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>SEED-Bench</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr style="background-color: #f8f9fa; font-weight: 600;">
<td colspan="10" class="has-text-centered" style="background-color: #e9ecef;">
<strong>Ophthalmology-Specific Benchmarks</strong>
</td>
</tr>
<tr>
<td><strong>Bench-Nephrology</strong></td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>Eval-GPT-Ophth</strong></td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>Bench-Myopia</strong></td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr>
<td><strong>OphNet</strong></td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✓</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
<td class="has-text-centered">✗</td>
</tr>
<tr style="background-color: #e8f5e8; font-weight: 600;">
<td><strong>LMOD+ (ours)</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
<td class="has-text-centered"><strong>✓</strong></td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">⚙️ Data Curation Pipeline</h2>
<div class="content has-text-justified">
<img src="./assets/webpage/static/images/data_pipeline.jpg" width="100%" />
<p>Our comprehensive data curation pipeline systematically processes ophthalmology datasets by extracting
key
clinical information, anatomical annotations, and diagnostic metadata. We leverage multimodal large
language
models to automatically generate diverse question-answer pairs spanning anatomical recognition, disease
diagnosis, staging assessment, and patient demographic analysis.</p>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">📊Statistics</h2>
<div class="content">
<table id="statistics-table" class="js-sort-table">
<thead>
<tr>
<th class="js-sort-string"><strong>Dataset</strong></th>
<th class="js-sort-string"><strong>Type</strong></th>
<th class="js-sort-number"><strong>Number</strong></th>
<th class="js-sort-string"><strong>Disease</strong></th>
</tr>
</thead>
<tbody>
<tr>
<td>Cataract-1K</td>
<td>Surgical Scene</td>
<td>2256</td>
<td>Cataract</td>
</tr>
<tr>
<td>Harvard FairSeg</td>
<td>Scanning Laser Ophthalmoscopy</td>
<td>10000</td>
<td>Retinal</td>
</tr>
<tr>
<td>OIMHS</td>
<td>Optical Coherence Tomography</td>
<td>3859</td>
<td>Macular</td>
</tr>
<tr>
<td>CAU001</td>
<td>Lens Photography</td>
<td>1417</td>
<td>N/A</td>
</tr>
<tr>
<td>Cataract Detection 2</td>
<td>Lens Photography</td>
<td>1015</td>
<td>Cataract</td>
</tr>
<tr>
<td>REFUGE</td>
<td>Color Fundus Photography</td>
<td>1200</td>
<td>Glaucoma</td>
</tr>
<tr>
<td>IDRiD</td>
<td>Color Fundus Photography</td>
<td>516</td>
<td>Retinopathy</td>
</tr>
<tr>
<td>ORIGA</td>
<td>Color Fundus Photography</td>
<td>650</td>
<td>Glaucoma</td>
</tr>
<tr>
<td>G1020</td>
<td>Color Fundus Photography</td>
<td>1020</td>
<td>Glaucoma</td>
</tr>
<tr>
<td>BRSET</td>
<td>Color Fundus Photography</td>
<td>16249</td>
<td>Multi</td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">🖼️ Data Samples</h2>
<div class="content has-text-justified">
<img src="./assets/webpage/static/images/eye_dataset.png" width="100%" />
<p>This image provides sample annotations from the eight ophthalmology datasets used in our experiments,
demonstrating the variety and complexity of regional markings. These datasets include: (a) Cataract-1K,
(b)
Harvard FairSeg, (c) OIMHS, (d) CAU001, (e) Cataract Detection 2, (f) REFUGE, (g) IDRiD, and (h) ORIGA.
</p>
</div>
</div>
</div>
</div>
</section>
<!-- RESULTS SECTION -->
<section class="hero is-info is-small"
style="background: linear-gradient(135deg, #74b9ff 0%, #0984e3 100%); margin-top: 2rem;">
<div class="hero-body has-text-centered">
<h1 class="title is-1 lmod">
<img src="./assets/webpage/static/images/goji_logo.png" style="width:1em;vertical-align: middle" alt="Logo" />
<span class="lmod" style="vertical-align: middle">Benchmark Results</span>
</h1>
</div>
</section>
<section class="section" style="padding-top: 3rem;">
<div class="container">
<!-------------------------------------------------------------------- RESULTS SECTION -------------------------------------------------------------------->
<!-- 1. Performance Comparison Across Imaging Modalities -->
<div class="columns is-centered m-3">
<div class="column is-full has-text-centered content">
<h2 class="title is-3">📊Ocular Anatomical Structure Recognition Performance</h2>
<div class="content">
<div class="content has-text-justified">
<img src="./assets/webpage/static/images/combined_radar_charts_horizontal.jpg" width="80%"
style="display: block; margin: 0 auto;" />
<p>
Comprehensive performance evaluation of leading multimodal large language models across diverse
ophthalmic imaging techniques.
The radar visualization illustrates how top-performing models achieve varying levels of precision,
recall, F1-score, and hit rate
across five critical imaging modalities in ophthalmology: surgical scenes, optical coherence tomography,
color fundus photography,
scanning laser ophthalmoscopy, and lens photography. This analysis reveals modality-specific strengths
and limitations of current
state-of-the-art models in medical image understanding.
</p>
</div>
</div>
</div>
</div>
<!-- 2. Binary Eye Condition Diagnosis -->
<div class="columns is-centered m-6">
<div class="column is-full has-text-centered content">
<h2 class="title is-3">🎯Binary Eye Condition Diagnosis Performance</h2>
<div class="content">
<div class="content has-text-justified">
<img src="./assets/webpage/static/images/binary_classification_heatmap.jpg" width="60%"
style="display: block; margin: 0 auto;" />
<p>
Comprehensive diagnostic accuracy assessment of 23 multimodal large language models across 12 distinct
eye conditions.
The heatmap visualization employs a color-coded accuracy scale from 0 to 1, where darker intensities
represent superior
diagnostic capabilities. This systematic evaluation demonstrates the varying proficiency of different
models in accurately
identifying specific ophthalmic conditions, highlighting both model-specific strengths and
disease-specific diagnostic challenges
in automated eye care diagnosis.
</p>
</div>
</div>
</div>
</div>
<!-- 3. Multi-class Eye Disease Diagnosis -->
<div class="columns is-centered m-6">
<div class="column is-full has-text-centered content">
<h2 class="title is-3">🔬Multi-class Eye Disease Classification Analysis</h2>
<div class="content">
<div class="content has-text-justified">
<img src="./assets/webpage/static/images/multiclass_diagnosis_scatter.jpg" width="70%"
style="display: block; margin: 0 auto;" />
<p>
Relationship analysis between model complexity and diagnostic performance in multi-class eye disease
classification using
color fundus photography. The scatter plot reveals how model parameter scale (measured in billions)
correlates with
diagnostic accuracy on a four-class classification task. Connected trajectories within model families
demonstrate
performance scaling patterns across different architectural configurations. The baseline random
performance threshold
(25% for four-class classification) is indicated by the gray dashed line, with selected LLaVA variants
specifically
labeled to highlight architectural distinctions and their impact on diagnostic capabilities.
</p>
</div>
</div>
</div>
</div>
<!-- 4. Stage-based Diagnosis Performance -->
<div class="columns is-centered m-6">
<div class="column is-full has-text-centered content">
<h2 class="title is-3">📈Ophthalmologic Stage Diagnosis Capabilities</h2>
<div class="content">
<div class="content has-text-justified">
<img src="./assets/webpage/static/images/stage_diagnosis_accuracy_chart.jpg" width="70%"
style="display: block; margin: 0 auto;" />
<p>
Comparative analysis of multimodal large language model performance on stage-based ophthalmologic
diagnosis tasks.
The bar chart evaluation encompasses 10 carefully selected models tested across three specialized
datasets requiring
precise stage classification: OIMHS Macular Hole staging, ICDR severity assessment, and SDRG grading.
Baseline
performance thresholds at 20% and 25% provide reference points for model evaluation. The analysis
includes diverse
model architectures from InternVL variants (ranging from 1.5-2B to 2.5-8B-MPO parameters), LLaVA family
models,
specialized LLaVA-Med-7B, QWen-7B, YI-VL-6B, and DeepSeek VL2-Tiny. Results demonstrate that ICDR tasks
achieve
the highest diagnostic accuracies (approaching 40%), while OIMHS Macular Hole staging and SDRG exhibit
more
consistent performance within the 15-25% range, with InternVL 2.5-8B showing particular excellence in
ICDR assessment.
</p>
</div>
</div>
</div>
</div>
<div class="columns is-centered m-6">
<div class="column is-full has-text-centered content">
<h2 class="title is-3" id="leaderboard">🏆LMOD+ Subset Leaderboard</h2>
<div class="content">
<div class="content has-text-justified">
<p>
We conduct comprehensive evaluation of state-of-the-art multimodal large language models (MLLMs) using
our proposed benchmarks, encompassing both proprietary and open-source models across multiple
performance
dimensions. Our evaluation framework focuses on three critical ophthalmological tasks: anatomical
recognition, disease diagnosis, and disease stage assessment. The LMOD+ Standard Leaderboard is
evaluated
on a carefully curated LMOD+ <a
href="https://yaleedu-my.sharepoint.com/:f:/r/personal/zhenyue_qin_yale_edu/Documents/Research/Data/LMOD/LMOD-Subset?csf=1&web=1&e=F5mZlx"
target="_blank" style="color: #2c5aa0; text-decoration: underline;">subset</a>, comprising 1,076
images specifically selected for
efficient and reliable model evaluation.
</p>
</div>
<!-- <button id="toggleButton" onclick="changeButtonText()"><b style='font-size: larger;'>Disease Diagnosis
Leaderboard</b> (Click to Switch)</button> -->
<div class="model-labels-container">
<!-- <span class="leaderboard-label" style="background-color: rgba(255, 208, 80, 0.15);">Human Expert</span> -->
<span class="leaderboard-label" style="background-color: rgba(249, 242, 248, 1);">Open-Source</span>
<span class="leaderboard-label" style="background-color: rgba(117, 209, 215, 0.1);">Proprietary</span>
</div>
<!-- Gold Standard Leaderboard -->
<table id="table1">
<thead>
<!-- First row: Main category headers -->
<tr style="background-color: #f8f9fa; border-bottom: 2px solid #dee2e6;">
<th rowspan="2" class="has-text-centered js-sort-none no-sort"
style="vertical-align: middle; border-right: 2px solid #dee2e6; padding: 12px 8px; font-weight: 700; color: #495057; cursor: default; pointer-events: none !important;">
<strong>Models</strong>
</th>
<th rowspan="2" class="has-text-centered js-sort-none no-sort"
style="vertical-align: middle; border-right: 2px solid #dee2e6; padding: 12px 8px; font-weight: 700; color: #495057; cursor: default; pointer-events: none !important;">
<strong>Release Date</strong>
</th>
<th colspan="4" class="has-text-centered"
style="border-right: 2px solid #dee2e6; padding: 12px 8px; font-weight: 700; color: #2c5aa0; background-color: #e8f4fd; cursor: default;">
<strong>Anatomical Recognition</strong>
</th>
<th colspan="2" class="has-text-centered"
style="border-right: 2px solid #dee2e6; padding: 12px 8px; font-weight: 700; color: #28a745; background-color: #e8f5e8; cursor: default;">
<strong>Diagnosis Analysis</strong>
</th>
<th rowspan="2" class="has-text-centered"
style="vertical-align: middle; padding: 12px 8px; font-weight: 700; color: #dc3545; background-color: #fde8e8;">
<strong>Staging Assessment<br />Acc</strong>
</th>
</tr>
<!-- Second row: Specific metric headers -->
<tr style="background-color: #f8f9fa; border-bottom: 1px solid #dee2e6;">
<th class="has-text-centered"
style="padding: 8px 6px; font-size: 13px; font-weight: 600; color: #2c5aa0; background-color: #e8f4fd;">
<strong>Prec.</strong>
</th>
<th class="has-text-centered"
style="padding: 8px 6px; font-size: 13px; font-weight: 600; color: #2c5aa0; background-color: #e8f4fd;">
<strong>Rec.</strong>
</th>
<th class="has-text-centered"
style="padding: 8px 6px; font-size: 13px; font-weight: 600; color: #2c5aa0; background-color: #e8f4fd;">
<strong>F1</strong>
</th>
<th class="has-text-centered"
style="padding: 8px 6px; font-size: 13px; font-weight: 600; color: #2c5aa0; background-color: #e8f4fd; border-right: 2px solid #dee2e6;">
<strong>HC</strong>
</th>
<th class="has-text-centered"
style="padding: 8px 6px; font-size: 13px; font-weight: 600; color: #28a745; background-color: #e8f5e8;">
<strong>Binary Acc</strong>
</th>
<th class="has-text-centered"
style="padding: 8px 6px; font-size: 13px; font-weight: 600; color: #28a745; background-color: #e8f5e8; border-right: 2px solid #dee2e6;">
<strong>Multi-class Acc</strong>
</th>
</tr>
</thead>
<tbody>
<!-- Random Baseline -->
<tr>
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">🎲 <b>Random</b></td>
<td style="text-align: center; padding: 10px 8px;"></td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">0.5000</td>
<td style="text-align: center; padding: 10px 8px;">0.2500</td>
<td style="text-align: center; padding: 10px 8px;">0.2500</td>
</tr>
<!-- Proprietary Models -->
<tr style="background-color: rgba(117, 209, 215, 0.1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
🤖 <b>GPT-4o</b>
</td>
<td style="text-align: center; padding: 10px 8px;">2024-05-13</td>
<td style="text-align: center; padding: 10px 8px;">0.3116</td>
<td style="text-align: center; padding: 10px 8px;">0.2424</td>
<td style="text-align: center; padding: 10px 8px;">0.2577</td>
<td style="text-align: center; padding: 10px 8px;">0.9786</td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">0.1053</td>
</tr>
<!-- Open-Source Models -->
<tr style="background-color: rgba(249, 242, 248, 1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
<img src="./assets/webpage/static/logo/llava_med_logo.png"
style="width: 20px; height: 20px; vertical-align: middle; margin-right: 8px;" alt="LLaVA Logo" />
<b>LLaVa-Med-v1.5-mistral-7B</b>
</td>
<td style="text-align: center; padding: 10px 8px;">2023-06-01</td>
<td style="text-align: center; padding: 10px 8px;">0.2090</td>
<td style="text-align: center; padding: 10px 8px;">0.2558</td>
<td style="text-align: center; padding: 10px 8px;">0.2098</td>
<td style="text-align: center; padding: 10px 8px;">0.6240</td>
<td style="text-align: center; padding: 10px 8px;">0.5000</td>
<td style="text-align: center; padding: 10px 8px;">0.2500</td>
<td style="text-align: center; padding: 10px 8px;">0.2368</td>
</tr>
<tr style="background-color: rgba(249, 242, 248, 1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
<img src="./assets/webpage/static/logo/Yi_logo.svg"
style="width: 20px; height: 20px; vertical-align: middle; margin-right: 8px;" alt="LLaVA Logo" />
<b>YI-VL-6B</b>
</td>
<td style="text-align: center; padding: 10px 8px;">2024-05-11</td>
<td style="text-align: center; padding: 10px 8px;">0.1509</td>
<td style="text-align: center; padding: 10px 8px;">0.0379</td>
<td style="text-align: center; padding: 10px 8px;">0.0451</td>
<td style="text-align: center; padding: 10px 8px;">0.8135</td>
<td style="text-align: center; padding: 10px 8px;">0.4117</td>
<td style="text-align: center; padding: 10px 8px;">0.2525</td>
<td style="text-align: center; padding: 10px 8px;"><b>0.4000</b></td>
</tr>
<tr style="background-color: rgba(249, 242, 248, 1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
<img src="./assets/webpage/static/logo/medflamingo_logo.webp"
style="width: 20px; height: 20px; vertical-align: middle; margin-right: 8px;" alt="LLaVA Logo" />
<b>Med-Flamingo</b>
</td>
<td style="text-align: center; padding: 10px 8px;">N/A</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
<td style="text-align: center; padding: 10px 8px;">INVALID</td>
</tr>
<tr style="background-color: rgba(249, 242, 248, 1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
<img src="./assets/webpage/static/logo/internVL_logo.webp"
style="width: 20px; height: 20px; vertical-align: middle; margin-right: 8px;" alt="LLaVA Logo" />
<b>InternVL 1.5-2B</b>
</td>
<td style="text-align: center; padding: 10px 8px;">2024-05-25</td>
<td style="text-align: center; padding: 10px 8px;">0.2377</td>
<td style="text-align: center; padding: 10px 8px;">0.2386</td>
<td style="text-align: center; padding: 10px 8px;">0.2204</td>
<td style="text-align: center; padding: 10px 8px;">0.7722</td>
<td style="text-align: center; padding: 10px 8px;">0.5000</td>
<td style="text-align: center; padding: 10px 8px;">0.2525</td>
<td style="text-align: center; padding: 10px 8px;">0.2632</td>
</tr>
<tr style="background-color: rgba(249, 242, 248, 1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
<img src="./assets/webpage/static/logo/internVL_logo.webp"
style="width: 20px; height: 20px; vertical-align: middle; margin-right: 8px;" alt="LLaVA Logo" />
<b>InternVL 1.5-4B</b>
</td>
<td style="text-align: center; padding: 10px 8px;">2024-05-25</td>
<td style="text-align: center; padding: 10px 8px;">0.2499</td>
<td style="text-align: center; padding: 10px 8px;">0.2541</td>
<td style="text-align: center; padding: 10px 8px;">0.2414</td>
<td style="text-align: center; padding: 10px 8px;">0.8096</td>
<td style="text-align: center; padding: 10px 8px;">0.5000</td>
<td style="text-align: center; padding: 10px 8px;">0.2500</td>
<td style="text-align: center; padding: 10px 8px;">0.2500</td>
</tr>
<tr style="background-color: rgba(249, 242, 248, 1);">
<td style="text-align: left; padding: 10px 8px; font-weight: 600;">
<img src="./assets/webpage/static/logo/internVL_logo.webp"
style="width: 20px; height: 20px; vertical-align: middle; margin-right: 8px;" alt="LLaVA Logo" />