ISIC 2019 Leaderboards
| Rank 64 total | Team 64 unique teams | Approach | Manuscript | Used External Data 19 yes | Primary Metric Value Balanced Accuracy | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 |
DAISYLab
Hamburg University of Technology/University Medical Center Hamburg-Eppendorf
|
Ensemble of Multi-Res EfficientNets + SEN154 2 | description |
public
Yes
|
0.636 | ||||||||||||||
|
|||||||||||||||||||
| 2 |
DysionAI
DYSION AI, Inc, Beijing, China
|
Ensemble of EfficienetB3-B4-Seresnext101 | description |
public_off
No
|
0.607 | ||||||||||||||
|
|||||||||||||||||||
| 3 |
AImageLab & PRHLT
Unimore & UPV
|
ensemble, ood threshold 100% | description |
public_off
No
|
0.593 | ||||||||||||||
|
|||||||||||||||||||
| 4 |
DermaCode
|
13 models + hierarchical approach to select outliers | description |
public_off
No
|
0.578 | ||||||||||||||
|
|||||||||||||||||||
| 5 |
Nurithm Labs
Nurithm Labs
|
Densenet-161 with heavy use of random crops | description |
public
Yes
|
0.569 | ||||||||||||||
|
|||||||||||||||||||
| 6 |
Torus Actions
Torus Actions
|
Simple test approach | description |
public_off
No
|
0.563 | ||||||||||||||
|
|||||||||||||||||||
| 7 |
BITDeeper
Beijing Institute of Technology
|
MelaNet: A Deep Dense Attention Network for Melanoma Detection in Dermoscopy Images | description |
public_off
No
|
0.558 | ||||||||||||||
|
|||||||||||||||||||
| 8 |
SYSU-MIA-Group
Sun Yat-sen University Medical Image Analyze Group
|
Class-centroid-Based Openset Classfication method on Skin Lession | description |
public_off
No
|
0.557 | ||||||||||||||
|
|||||||||||||||||||
| 9 |
MelanoNorm_IITRopar
Indian Institute of Technology Ropar
|
Classification Using Stacking and Long-Tail Distribution: High Confidence Interval | description |
public_off
No
|
0.546 | ||||||||||||||
|
|||||||||||||||||||
| 10 |
MH_team
|
Softmax Ensemble Model and Sigmoid Ensemble Model | description |
public_off
No
|
0.544 | ||||||||||||||
|
|||||||||||||||||||
| 11 |
BGU_hackers
Ben Gurion University
|
AAAR Ensemble pruning Approach | description |
public
Yes
|
0.543 | ||||||||||||||
|
|||||||||||||||||||
| 12 |
SabanciUnivTeam
Sabanci University
|
Submission 3 - Anomaly Detection | description |
public_off
No
|
0.533 | ||||||||||||||
|
|||||||||||||||||||
| 13 |
deltamicro
Delta Micro Technology Inc
|
transferlearning-ensemble-averaging | description |
public
Yes
|
0.532 | ||||||||||||||
|
|||||||||||||||||||
| 14 |
offer_show
iSee-SYSU
|
Ensemble of model 1 and model 2 | description |
public_off
No
|
0.532 | ||||||||||||||
|
|||||||||||||||||||
| 15 |
MIP
Southern Medical University
|
ensemble three ResNext50 with SE Block | description |
public_off
No
|
0.531 | ||||||||||||||
|
|||||||||||||||||||
| 16 |
Tencent Medical AI Lab
|
Top models ensemble with threshold | description |
public_off
No
|
0.525 | ||||||||||||||
|
|||||||||||||||||||
| 17 |
bashiri
|
Deep Convolution Neural Network with data augmentation | description |
public_off
No
|
0.523 | ||||||||||||||
|
|||||||||||||||||||
| 18 |
Airdoctor
|
Embedding vectors and Ensemble models | description |
public
Yes
|
0.522 | ||||||||||||||
|
|||||||||||||||||||
| 19 |
One-Two-Three
|
VGG-16 for Skin Lesion Diagnosis | description |
public_off
No
|
0.519 | ||||||||||||||
|
|||||||||||||||||||
| 20 |
shallow learning
Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
|
Skin Lesion Analysis Towards Melanoma Detection using Deep Neural Networks | description |
public_off
No
|
0.518 | ||||||||||||||
|
|||||||||||||||||||
| 21 |
VisinVis
ETRI
|
Ensembled Transfer Neural Networks by using Lesion Correlation Learning : Approach 3 | description |
public_off
No
|
0.513 | ||||||||||||||
|
|||||||||||||||||||
| 22 |
cu
cuhk
|
Ensembles with external data | description |
public
Yes
|
0.510 | ||||||||||||||
|
|||||||||||||||||||
| 23 |
CureSkinAI
CureSkin
|
Convolutional Neural Networks Ensemble towards Skin Lesion Analysis of Dermoscopic Images | description |
public_off
No
|
0.509 | ||||||||||||||
|
|||||||||||||||||||
| 24 |
Predicthy LLC
Predicthy LLC
|
Take max value among the combination of four neural networks | description |
public
Yes
|
0.507 | ||||||||||||||
|
|||||||||||||||||||
| 25 |
Hsinwei
|
Convolutional Ensemble with Out-of-Distribution Detector | description |
public
Yes
|
0.505 | ||||||||||||||
|
|||||||||||||||||||
| 26 |
MMU-VCLab
Manchester Metropolitan University
|
Two-stage Ensemble Method | description |
public
Yes
|
0.502 | ||||||||||||||
|
|||||||||||||||||||
| 27 |
Yongsheng Pan
MICAL
|
FV-RES | description |
public_off
No
|
0.501 | ||||||||||||||
|
|||||||||||||||||||
| 28 |
Pan Galactic Gargleblasters test
|
Deep Learning Resnet50 | description |
public_off
No
|
0.493 | ||||||||||||||
|
|||||||||||||||||||
| 29 |
SY2
Beihang University
|
Skin Lesion Diagnosis using color constancy and loss weighting with external dataset 03 | description |
public
Yes
|
0.492 | ||||||||||||||
|
|||||||||||||||||||
| 30 |
logreg
|
EfficientNet b1 with augmentations | description |
public_off
No
|
0.492 | ||||||||||||||
|
|||||||||||||||||||
| 31 |
MGI
National Institutes of Biotechnology Malaysia
|
Densenet161 with discriminative learning rate | description |
public_off
No
|
0.489 | ||||||||||||||
|
|||||||||||||||||||
| 32 |
gxl_xgy_llz_victory
Institute of automation, Chinese academy of sciences
|
Ensembled Model for Skin Lesion Classification | description |
public_off
No
|
0.489 | ||||||||||||||
|
|||||||||||||||||||
| 33 |
Mt.Smart
MTlab,Meitu Inc
|
Multiple Convolution Neural Net Ensemble | description |
public_off
No
|
0.482 | ||||||||||||||
|
|||||||||||||||||||
| 34 |
BMIT
Biomedical and Multimedia Information Technology, University of Sydney
|
152 Layer Resnet | description |
public_off
No
|
0.481 | ||||||||||||||
|
|||||||||||||||||||
| 35 |
Aiden Gatani
|
densenet201 | description |
public_off
No
|
0.470 | ||||||||||||||
|
|||||||||||||||||||
| 36 |
Le-Health
Lenovo Research
|
ensemble strategy 2 | description |
public_off
No
|
0.469 | ||||||||||||||
|
|||||||||||||||||||
| 37 |
SRMC
DataGenius
|
Approch 1 : Usable Predictive Model - Densenet121 | description |
public
Yes
|
0.465 | ||||||||||||||
|
|||||||||||||||||||
| 38 |
SY1
Beihang university
|
Skin Lesion Diagnosis using discrimination criterion | description |
public_off
No
|
0.464 | ||||||||||||||
|
|||||||||||||||||||
| 39 |
Panetta's Vision and Sensing Systems Lab
Tufts University
|
XGBoost Ensemble | description |
public
Yes
|
0.461 | ||||||||||||||
|
|||||||||||||||||||
| 40 |
IML group - DFKI
Interactive Machine Learning (IML) - German Research Center for Artificial Intelligence (DFKI)
|
One-class SVM pre-filter + VGG16 CNN | description |
public_off
No
|
0.445 | ||||||||||||||
|
|||||||||||||||||||
| 41 |
SIGMA
NUST
|
Ensemble of Fine-tuned DNNs for Skin Lesion Image Classi cation | description |
public
Yes
|
0.438 | ||||||||||||||
|
|||||||||||||||||||
| 42 |
skychain
skychain
|
resnext50_32x4d | description |
public_off
No
|
0.437 | ||||||||||||||
|
|||||||||||||||||||
| 43 |
SJ_T1
|
Ten1 | description |
public
Yes
|
0.429 | ||||||||||||||
|
|||||||||||||||||||
| 44 |
KDIS
University of Cordoba & Maimonides Biomedical Research Institute of Cordoba
|
Multi-view convolutional architecture - Margin sampling version | description |
public_off
No
|
0.429 | ||||||||||||||
|
|||||||||||||||||||
| 45 |
SUMMER
|
Feature Fusion for Accurate Skin Lesion Analysis | description |
public
Yes
|
0.427 | ||||||||||||||
|
|||||||||||||||||||
| 46 |
AIRL
Central South University
|
Skin Lesion Classification with Out-of-Distribution Detection Using Deep Neural Network Ensemble | description |
public_off
No
|
0.426 | ||||||||||||||
|
|||||||||||||||||||
| 47 |
sysutest1
sysu
|
mcd1 | description |
public_off
No
|
0.423 | ||||||||||||||
|
|||||||||||||||||||
| 48 |
I don't know what to eat
|
SEnet154 | description |
public_off
No
|
0.419 | ||||||||||||||
|
|||||||||||||||||||
| 49 |
UH ML Lab
University of Hawaii at Manoa
|
Inceptionv3 with CBAM | description |
public_off
No
|
0.419 | ||||||||||||||
|
|||||||||||||||||||
| 50 |
LLCW
Institute of Automation, Chinese Academy of Sciences
|
InceptionV3 with Transfer Learning | description |
public
Yes
|
0.405 | ||||||||||||||
|
|||||||||||||||||||
| 51 |
sstl
上海计算机软件技术开发中心
|
Fine-grained skin image classification | description |
public
Yes
|
0.370 | ||||||||||||||
|
|||||||||||||||||||
| 52 |
Cihan Soylu
|
Transfer learning with DenseNet201 | description |
public_off
No
|
0.370 | ||||||||||||||
|
|||||||||||||||||||
| 53 |
CrazyLearningTeam
|
CrazyLearningTeam | description |
public_off
No
|
0.346 | ||||||||||||||
|
|||||||||||||||||||
| 54 |
SB
EPFL
|
Skin Lesion Analysis Towards Melanoma Detection using Siamese neural network | description |
public_off
No
|
0.329 | ||||||||||||||
|
|||||||||||||||||||
| 55 |
SIBET CAS
Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences
|
Semi-Supervised GAN | description |
public_off
No
|
0.327 | ||||||||||||||
|
|||||||||||||||||||
| 56 |
UTHealth-Onto.
University of Texas Health Science Center at Houston
|
Random Forest Ensemble_xy | description |
public_off
No
|
0.315 | ||||||||||||||
|
|||||||||||||||||||
| 57 |
SharpestMinds
SharpestMinds
|
Data Augmentation and Transfer Learning with ResNet50 with Cyclical Learning Rates | description |
public_off
No
|
0.304 | ||||||||||||||
|
|||||||||||||||||||
| 58 |
SkinLegion
Persistent Systems Ltd
|
Baseline Classifier | description |
public_off
No
|
0.273 | ||||||||||||||
|
|||||||||||||||||||
| 59 |
mvlab-skin
Indian Institute of Technology Roorkee
|
ISIC 2019 : Deep Ensembled Framework for Skin Lesion Analysis Towards Melanoma Detection (DEFSMD) | description |
public
Yes
|
0.258 | ||||||||||||||
|
|||||||||||||||||||
| 60 |
YouMe AI
|
Ensembles of Deep Convolution Neural Networks - updated | description |
public_off
No
|
0.251 | ||||||||||||||
|
|||||||||||||||||||
| 61 |
IT Derm Lab
DBE
|
Fine tune Resnet | description |
public_off
No
|
0.109 | ||||||||||||||
|
|||||||||||||||||||
| 62 |
jasdeep
Singh
|
Asymmetrical loss function and cutout along the corners | description |
public
Yes
|
0.108 | ||||||||||||||
|
|||||||||||||||||||
| 63 |
TUKL
NUST
|
ResNet50 ConvNet using cyclical learning rates and transfer learning | description |
public_off
No
|
0.106 | ||||||||||||||
|
|||||||||||||||||||
| 64 |
P
|
ResNet50 | description |
public_off
No
|
0.048 | ||||||||||||||
|
|||||||||||||||||||