MILK10k 2025 Leaderboards

Lesion Diagnosis
Rank 76 total Team 76 unique teams Approach Manuscript Used External Data 14 yes Primary Metric Value Dice Coefficient
1
Genial Team
 Medgemma (Medsiglip) v3 -
public Yes
0.683
2
Trần Văn Đồng - Beast Team
Independant researchers
EfficientnetV2 + customized loss v9 add milk10 -
public Yes
0.654
3
Rohan Iyer
University of Minnesota
F1-Weighted Tuned Ensemble VI (ConvNeXt + Swin-V2) -
public_off No
0.609
4
Rohan Iyer + External Data
University of Minnesota
Dual Ensemble I -
public Yes
0.602
5
Sébastien Fischman
Damae Medical
Dermo only CNN updated -
public Yes
0.595
6
ENS
 Multi-Modal Swin V4 PT SM LA -
public Yes
0.560
7
DermaNet
DermaNet Team
DermaNet - No External -
public_off No
0.549
8
Sabancı University ENS491
 rare classes swin v2 -
public Yes
0.548
9
Sabancı University ENS491 Team
 rare classes swin -
public Yes
0.546
10
STRAWB_MILK
QMUL
ViT_ConvNext_Ensemble_v1.4 -
public Yes
0.538
11
AcuvizionAI Belgium
Independent Researcher (Kamran Soleimani)
AcuvizionAI Multi-Modal Fusion -
public_off No
0.538
12
Multimodale
 Multimodal -
public_off No
0.532
13
PanDerm Team
AIM for Health Lab, Monash University
PanDerm V2 -
public_off No
0.524
14
Benjamin Bruch
Independent Researcher
Hybrid ViT-CNN Ensemble -
public Yes
0.522
15
Hongyuan Xie
Manchester Metropolitan University
Avg_Scale_Size_CNN -
public_off No
0.519
16
SkinVP
Sabanci university
Efficient_net_train_bothRare -
public Yes
0.506
17
PikaPika
KAIST
EfficientnetV2 + Metadata MLP Fusion -
public_off No
0.505
18
PawsMakeClickClick
Kharkiv National University of Radio Electronics
Derm_v2 -
public_off No
0.504
19
able_23
PEC Chandigarh
convnext_base + asl -
public_off No
0.499
20
JTeam
 dual_vision -
public_off No
0.495
21
Valor
BRAC University b
ConvNeXt +SwinV2 -
public_off No
0.495
22
BUPT
Beijing University of Posts and Telecommunications
MDCL-fold5 -
public_off No
0.491
23
Unknown
Hidden
Train 2 -
public_off No
0.489
24
Pedro Bento
 SwinV2 (tiny) - 2 epoch mid training -
public Yes
0.486
25
DL_2025.1_HUST
 PanDerm Tri-Modal Fusion -
public_off No
0.486
26
MDCL-SCRL-2
 MDCL-SCRL-FOLD3 -
public_off No
0.485
27
Bajnok
 Milk10k swin and effnet -
public_off No
0.482
28
TmpMusty
Independent Researcher
EfficientNetV2 multimodal dumen -
public_off No
0.481
29
ManuelPla
UOC
Multimodal 1.9 -
public_off No
0.478
30
ElaheVisionX SkinAnalytics
 EffNetV2 -
public_off No
0.477
31
LIFE
Universidade Federal do Espírito Santo
Concatenation baseline custom loss -
public Yes
0.476
32
milkyway
Yonsei iie4123
sampler+customlossv4+v1 ensemble -
public_off No
0.473
33
SzaboAndrasBajnok
University of Szeged
Multimodal -
public_off No
0.472
34
MDCL-SCRL
 MDCL-SCRL-FOLD1 -
public_off No
0.471
35
MDCL-SCRL-all
 ALL -
public_off No
0.471
36
Nikshay Kataria
 convnext_512 -
public_off No
0.470
37
Rekura
 Concept bottleneck model for skin lesion classification description
public_off No
0.469
38
SkinGuard
 ResNet50 -
public_off No
0.464
39
Huvy
 Volo D1 Remix -
public Yes
0.463
40
AK
HUST
CNN -
public_off No
0.462
41
Judy
BUAA
resnet50 -
public_off No
0.461
42
It Hurts When IP
Woxsen University
Densenet -
public_off No
0.459
43
SZTE
 effnet_multi -
public_off No
0.457
44
Unknown2
Unknown
One fold -
public_off No
0.457
45
ZY
AHU
Dual_SwinTransformer -
public_off No
0.456
46
cheesenet
 Efficientnetv2 + custom loss + dynamics -
public_off No
0.447
47
Dhruval Padia
 Ensemble -
public_off No
0.444
48
John's Lab
Independent
Vision transformer - custom head - 6 -
public_off No
0.444
49
NIH NLM AI4M
National Institutes of Health, National Library of Medicine
Multimodal-Image-Text -
public_off No
0.442
50
CBML_UNIPI
University of Pireaus
EfficientNetV2S_DualFusion_Attention -
public_off No
0.439
51
Erdem Bozyel
 Ensemble v1 -
public Yes
0.437
52
Concat-four
 Concat-fold5 -
public_off No
0.436
53
Concat
 Concat-All -
public_off No
0.424
54
DBAA
Hanoi University of Science and Technology
AMMFA -
public_off No
0.419
55
AIEM
Johns Hopkins University
CNN -
public_off No
0.419
56
ZCU-FAV-KKY
University of West Bohemia
CAFormer18 -
public_off No
0.415
57
Soland
UET
EfficientNetV2 -
public_off No
0.413
58
ENS-Group#2
Sabanci University
EfficientNetV2-S Multi-Label Dermoscopic Classifier -
public_off No
0.413
59
NoExternalData
 NoExternalData used -
public_off No
0.410
60
SophiaMSCore_Systems
 ConvNeXt +SwinV2 -
public_off No
0.405
61
BUPT6
Beijing University of Posts and Telecommunications
ResNext-Pre-6 -
public_off No
0.402
62
XRAI-Keele
Keele University
Swin-v2 with concept-guided cross-attention -
public_off No
0.389
63
slemanfaiz
 ResNet optimized -
public_off No
0.384
64
ainum
 CLIP pretrained encoders with cross attention -
public_off No
0.366
65
Milk10k
 Swin v1 -
public_off No
0.353
66
David Nduonofit
Bowen University
EfficientNetV2S Multimodal fusion -
public_off No
0.346
67
Yhltmxh
TEMP
EfficientnetV2 -
public_off No
0.328
68
BUPT3
Beijing University of Posts and Telecommunications
aa2 -
public_off No
0.323
69
BUPT5
Beijing University of Posts and Telecommunications
ResNext-OnlyImage -
public_off No
0.317
70
TestTeam
 TestModel -
public_off No
0.296
71
JMI SkinVision (Farhan khan)
Jamia Millia Islamia
Dual-EfficientNetV2 with Multi Modal Fusion -
public_off No
0.260
72
BUPT4
Beijing University of Posts and Telecommunications
ResNext-OnlyImage -
public_off No
0.239
73
skintell
 SkinTell -
public_off No
0.236
74
catalinapesquet
 MobileNet_clinical -
public_off No
0.200
75
fibers
Universitat Politècnica de Barcelona
First test -
public_off No
0.136
76
KW_HAU
Kwangwoon University
Test -
public_off No
0.000