ISIC 2018 Leaderboards
| Rank 77 total | Team 77 unique teams | Approach | Manuscript | Used External Data 19 yes | Primary Metric Value Balanced Accuracy | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 |
MetaOptima Technology Inc.
MetaOptima Technology Inc.
|
Top 10 Models Averaged | description |
public
Yes
|
0.885 | ||||||||||||||
|
|||||||||||||||||||
| 2 |
DAISYLab
DAISYLab
|
Large Ensemble with heavy multi-cropping and loss weighting | description |
public
Yes
|
0.856 | ||||||||||||||
|
|||||||||||||||||||
| 3 |
Medical Image Analysis Group, Sun Yat-sen University
Medical Image Analysis Group, Sun Yat-sen University
|
Emsemble Of SENET and PNANET with DataAugmentation when TEST | description |
public_off
No
|
0.845 | ||||||||||||||
|
|||||||||||||||||||
| 4 |
Li
Li
|
densenet | description |
public_off
No
|
0.815 | ||||||||||||||
|
|||||||||||||||||||
| 5 |
Ask Sina
Ask Sina
|
Approach 3 : Average of Approach 1 and 2 | description |
public_off
No
|
0.812 | ||||||||||||||
|
|||||||||||||||||||
| 6 |
RECOD Titans
RECOD Titans
|
Average of 15 Deep Learning Models Trained Only with Challenge Data | description |
public_off
No
|
0.803 | ||||||||||||||
|
|||||||||||||||||||
| 7 |
NWPU-SAIIP
NWPU-SAIIP
|
FV+Res101 | description |
public_off
No
|
0.786 | ||||||||||||||
|
|||||||||||||||||||
| 8 |
Wonlab in Sungkyunkwan University, Korea, Republic of
Wonlab in Sungkyunkwan University, Korea, Republic of
|
WonDerM: Skin Lesion Classification with Fine-tuned Neural Networks | description |
public
Yes
|
0.785 | ||||||||||||||
|
|||||||||||||||||||
| 9 |
vess
vess
|
Resnext101 & DPN92, Snapshot ensamble, D4 TTA | description |
public_off
No
|
0.784 | ||||||||||||||
|
|||||||||||||||||||
| 10 |
LMU DataMining
LMU DataMining
|
thresholding DF AKIEC MEL VASC BKL | description |
public
Yes
|
0.780 | ||||||||||||||
|
|||||||||||||||||||
| 11 |
University of Washington
University of Washington
|
Densenet201 | description |
public_off
No
|
0.768 | ||||||||||||||
|
|||||||||||||||||||
| 12 |
Holidayburned
Holidayburned
|
Ensemble_of_resnet_and_inception_iteration_20000 | description |
public
Yes
|
0.754 | ||||||||||||||
|
|||||||||||||||||||
| 13 |
DeepOncology.AI
DeepOncology.AI
|
DeepOncology.AI-real_test_0.915_ensemble | description |
public_off
No
|
0.753 | ||||||||||||||
|
|||||||||||||||||||
| 14 |
University of Minnesota Center for Distributed Robotics
University of Minnesota Center for Distributed Robotics
|
An ensemble of resnet, densenet, inception, xception, and inceptionresnet | description |
public_off
No
|
0.752 | ||||||||||||||
|
|||||||||||||||||||
| 15 |
BIL, NTU
BIL, NTU
|
Conv. Ensemble (Inception Models Normalized+Un-Norm) + avg. pooling | description |
public_off
No
|
0.736 | ||||||||||||||
|
|||||||||||||||||||
| 16 |
Tufts University School of Medicine
Tufts University School of Medicine
|
Convnet ResNet50 with cyclical learning rates V2 | description |
public_off
No
|
0.735 | ||||||||||||||
|
|||||||||||||||||||
| 17 |
Le-Health
Le-Health
|
Classification Using the Cascade Structure | description |
public_off
No
|
0.734 | ||||||||||||||
|
|||||||||||||||||||
| 18 |
350818
350818
|
multidimensional ensembling | description |
public_off
No
|
0.728 | ||||||||||||||
|
|||||||||||||||||||
| 19 |
Nile University - MIIP
Nile University - MIIP
|
Ensembles of ResNet50, ResNet34, InceptionV3 | description |
public_off
No
|
0.728 | ||||||||||||||
|
|||||||||||||||||||
| 20 |
QuindiTech: Xuan Li, Peng Xu
QuindiTech: Xuan Li, Peng Xu
|
Inception Ensemble | description |
public_off
No
|
0.727 | ||||||||||||||
|
|||||||||||||||||||
| 21 |
Tandon Titans
Tandon Titans
|
ResNet-50-Post-Processing | description |
public
Yes
|
0.725 | ||||||||||||||
|
|||||||||||||||||||
| 22 |
Texas A&M Aggies
Texas A&M Aggies
|
Sequential PNASNet Classification based on Balanced Color-Normed Dataset | description |
public_off
No
|
0.724 | ||||||||||||||
|
|||||||||||||||||||
| 23 |
Hangzhou Dianzi University CAD429
Hangzhou Dianzi University CAD429
|
CNN; ensemble learning; multi-features. | description |
public_off
No
|
0.721 | ||||||||||||||
|
|||||||||||||||||||
| 24 |
2nd Appinion
2nd Appinion
|
Wide residual network applied to 7-class skin lesion classification | description |
public_off
No
|
0.721 | ||||||||||||||
|
|||||||||||||||||||
| 25 |
Hdu CAD429
Hdu CAD429
|
Ensemble with many Multi Scale Convolutional Neural Network | description |
public_off
No
|
0.718 | ||||||||||||||
|
|||||||||||||||||||
| 26 |
Dysion AI Technology Co., Ltd
Dysion AI Technology Co., Ltd
|
Deep Model Ensemble with Data Alignment v1 | description |
public
Yes
|
0.716 | ||||||||||||||
|
|||||||||||||||||||
| 27 |
AI Toulouse
AI Toulouse
|
Using Znet and bagging from two first approachs | description |
public
Yes
|
0.715 | ||||||||||||||
|
|||||||||||||||||||
| 28 |
Department of Dermatology, University of Rzeszów, Poland
Department of Dermatology, University of Rzeszów, Poland
|
ResNet101-SGD | description |
public_off
No
|
0.709 | ||||||||||||||
|
|||||||||||||||||||
| 29 |
QuindiTech: Yuchen Lu
QuindiTech: Yuchen Lu
|
ls + resnet + ens | description |
public_off
No
|
0.708 | ||||||||||||||
|
|||||||||||||||||||
| 30 |
BioImaging-KHU
BioImaging-KHU
|
Deep Learning with Adapted InceptionResNetV2 | description |
public_off
No
|
0.705 | ||||||||||||||
|
|||||||||||||||||||
| 31 |
Dominiks AI team
Dominiks AI team
|
Above dermatologist-level classification of malignant melanomas with deep neural | description |
public
Yes
|
0.703 | ||||||||||||||
|
|||||||||||||||||||
| 32 |
Mammoth
Mammoth
|
Old fashion | description |
public_off
No
|
0.703 | ||||||||||||||
|
|||||||||||||||||||
| 33 |
Redha Ali, Russell C. Hardie, Manawaduge Supun De Silva, and Temesguen Messay Ke
Redha Ali, Russell C. Hardie, Manawaduge Supun De Silva, and Temesguen Messay Ke
|
Combining Deep and Handcrafted Image Features for Skin Cancer Classification | description |
public_off
No
|
0.701 | ||||||||||||||
|
|||||||||||||||||||
| 34 |
CNR-ISASI_Lecce
CNR-ISASI_Lecce
|
Deep Convolutional Neural Network with Stochastic Gradient Descent Optimization | description |
public_off
No
|
0.699 | ||||||||||||||
|
|||||||||||||||||||
| 35 |
Hosei University, Iyatomi lab
Hosei University, Iyatomi lab
|
SEResNet101 w/ mean_teacher + SEResNet152 w/o mean_teacher | description |
public_off
No
|
0.695 | ||||||||||||||
|
|||||||||||||||||||
| 36 |
Manu Goyal
Manu Goyal
|
DeeplabV3+ with Priority strategy based on benign/maligant and number of imag | description |
public_off
No
|
0.695 | ||||||||||||||
|
|||||||||||||||||||
| 37 |
PA_Tech
PA_Tech
|
deep convolutional neural network with transfer learning | description |
public_off
No
|
0.692 | ||||||||||||||
|
|||||||||||||||||||
| 38 |
Opsins
Opsins
|
Transfer learning based CNN | description |
public_off
No
|
0.691 | ||||||||||||||
|
|||||||||||||||||||
| 39 |
UNIST_BMIPL
UNIST_BMIPL
|
Multiscale Lesion Segmentation and Application to Skin Cancer Classification | description |
public_off
No
|
0.687 | ||||||||||||||
|
|||||||||||||||||||
| 40 |
Nitwit AI
Nitwit AI
|
Inception V3 | description |
public_off
No
|
0.685 | ||||||||||||||
|
|||||||||||||||||||
| 41 |
The Homeboy's
The Homeboy's
|
Fusion of classical and DL technique | description |
public_off
No
|
0.684 | ||||||||||||||
|
|||||||||||||||||||
| 42 |
Tencent Youtu Lab
Tencent Youtu Lab
|
ResNet based skin lesion diagnosis with triplet loss and feature similarity | description |
public
Yes
|
0.683 | ||||||||||||||
|
|||||||||||||||||||
| 43 |
Persistent Systems
Persistent Systems
|
Two-stage hierarchical classifier | description |
public_off
No
|
0.680 | ||||||||||||||
|
|||||||||||||||||||
| 44 |
DC
DC
|
resnet ensemble | description |
public_off
No
|
0.679 | ||||||||||||||
|
|||||||||||||||||||
| 45 |
CNB-CSIC & FTN-UNS
CNB-CSIC & FTN-UNS
|
Ensemble of transfer learning on VGG16 and GoogLeNet | description |
public_off
No
|
0.676 | ||||||||||||||
|
|||||||||||||||||||
| 46 |
Università degli Studi di Modena e Reggio Emilia (AImage Lab .zip)
Università degli Studi di Modena e Reggio Emilia (AImage Lab .zip)
|
inception fine-tuned, weighted loss and data augmentation | description |
public_off
No
|
0.676 | ||||||||||||||
|
|||||||||||||||||||
| 47 |
WVUmich
WVUmich
|
resenet without augmentation fusion epochs | description |
public
Yes
|
0.668 | ||||||||||||||
|
|||||||||||||||||||
| 48 |
Computer Vision Lab, Nankai University
Computer Vision Lab, Nankai University
|
CNN ensemble | description |
public
Yes
|
0.665 | ||||||||||||||
|
|||||||||||||||||||
| 49 |
Stony Brook University
Stony Brook University
|
Transfer Learning 2 (0.9) | description |
public_off
No
|
0.658 | ||||||||||||||
|
|||||||||||||||||||
| 50 |
ISI_NLP_LAB
ISI_NLP_LAB
|
Ensembel of ResNet50, DenseNet121 and MobileNet | description |
public_off
No
|
0.657 | ||||||||||||||
|
|||||||||||||||||||
| 51 |
SAIIP-MIA
SAIIP-MIA
|
A Multi-Level Deep Ensemble Model | description |
public_off
No
|
0.655 | ||||||||||||||
|
|||||||||||||||||||
| 52 |
University of Padua
University of Padua
|
Ensemble of different CNN topologies | description |
public_off
No
|
0.651 | ||||||||||||||
|
|||||||||||||||||||
| 53 |
UnB
UnB
|
Task1 preprocessed K-fold Ensemble of Resnet50 | description |
public
Yes
|
0.650 | ||||||||||||||
|
|||||||||||||||||||
| 54 |
LTS5
LTS5
|
Convolutional Neural Network, DermoNet segmentation, ResNet50 | description |
public_off
No
|
0.650 | ||||||||||||||
|
|||||||||||||||||||
| 55 |
CAMP TUM
CAMP TUM
|
Webly Supervised Learning for Skin Lesion Classification | description |
public
Yes
|
0.639 | ||||||||||||||
|
|||||||||||||||||||
| 56 |
miltonbd
miltonbd
|
SENet-154 final valid 0.77 | description |
public
Yes
|
0.638 | ||||||||||||||
|
|||||||||||||||||||
| 57 |
Mehdi&peyman
Mehdi&peyman
|
InceptionV3 with augmentation | description |
public_off
No
|
0.603 | ||||||||||||||
|
|||||||||||||||||||
| 58 |
gchhor
gchhor
|
Pre-trained Inception-v3 | description |
public
Yes
|
0.601 | ||||||||||||||
|
|||||||||||||||||||
| 59 |
University of Dayton, Signal and Image Processing Lab
University of Dayton, Signal and Image Processing Lab
|
SVM classifier with hand-crafted features using segmentation | description |
public_off
No
|
0.577 | ||||||||||||||
|
|||||||||||||||||||
| 60 |
kevint
kevint
|
Two-Step Training of Deep Residual Networks for Skin Lesion Diagnosis | description |
public_off
No
|
0.572 | ||||||||||||||
|
|||||||||||||||||||
| 61 |
Andrey Sorokin
Andrey Sorokin
|
Hybrid Model of Lesion Boundary Detector and lesion Pixel-wise class detection | description |
public_off
No
|
0.563 | ||||||||||||||
|
|||||||||||||||||||
| 62 |
Andreas Pirchner
Andreas Pirchner
|
Generative adversarial networks for skin lesion classification | description |
public_off
No
|
0.515 | ||||||||||||||
|
|||||||||||||||||||
| 63 |
Team MLMI
Team MLMI
|
Cascaded DenseNets for Multi-Class Skin Lesion Classification | description |
public_off
No
|
0.501 | ||||||||||||||
|
|||||||||||||||||||
| 64 |
NTHU CVLab
NTHU CVLab
|
Deep learning with ResNet-50 | description |
public_off
No
|
0.499 | ||||||||||||||
|
|||||||||||||||||||
| 65 |
University of York
University of York
|
Multitask learning, single CNN for all three tasks, segmentation via FCN | description |
public
Yes
|
0.496 | ||||||||||||||
|
|||||||||||||||||||
| 66 |
SSNMLRG
SSNMLRG
|
hierarchy_resnet | description |
public_off
No
|
0.484 | ||||||||||||||
|
|||||||||||||||||||
| 67 |
QuindiTech: Pingao Wang
QuindiTech: Pingao Wang
|
Ensemble of Transfer Learning with pre-trained model over collected ISIC Archive | description |
public
Yes
|
0.463 | ||||||||||||||
|
|||||||||||||||||||
| 68 |
QuindiTech: Pingao Wang
QuindiTech: Pingao Wang
|
Transfer Learning with ISIC2018 archive | description |
public
Yes
|
0.449 | ||||||||||||||
|
|||||||||||||||||||
| 69 |
Balazs Harangi
Balazs Harangi
|
SKIN LESION DETECTION BASED ON AN ENSEMBLE OF ONE VERSUS ALL TRAINED CNN | description |
public_off
No
|
0.447 | ||||||||||||||
|
|||||||||||||||||||
| 70 |
LABCIN
LABCIN
|
Handcrafted features based on ABCD rule and ELM Hierarchical Classification | description |
public_off
No
|
0.445 | ||||||||||||||
|
|||||||||||||||||||
| 71 |
University of Illinois Springfield
University of Illinois Springfield
|
Multiple CNN model | description |
public_off
No
|
0.439 | ||||||||||||||
|
|||||||||||||||||||
| 72 |
Altumview and UBC Joint Team
Altumview and UBC Joint Team
|
Automatic Skin Lesion Analysis using Sliced Dermoscopy Images and Deep Residual | description |
public_off
No
|
0.424 | ||||||||||||||
|
|||||||||||||||||||
| 73 |
Deep-Class
Deep-Class
|
CNNCBR | description |
public_off
No
|
0.413 | ||||||||||||||
|
|||||||||||||||||||
| 74 |
Math & Stat Dept., UNC-Greensboro, USA and CSIE Dept., NTNU, Taiwan
Math & Stat Dept., UNC-Greensboro, USA and CSIE Dept., NTNU, Taiwan
|
Support vector machine with topological features from persistent homology III | description |
public_off
No
|
0.394 | ||||||||||||||
|
|||||||||||||||||||
| 75 |
University of Texas Freshman Research Initiative
University of Texas Freshman Research Initiative
|
Deep Learning with augmentation, dropout, and svm | description |
public_off
No
|
0.385 | ||||||||||||||
|
|||||||||||||||||||
| 76 |
SSTL
SSTL
|
Skin Disease Classification by Resnet50 | description |
public_off
No
|
0.360 | ||||||||||||||
|
|||||||||||||||||||
| 77 |
Michigan
Michigan
|
weighted super learner | description |
public_off
No
|
0.301 | ||||||||||||||
|
|||||||||||||||||||