top of page
MicrosoftTeams-image.jpg

Public International Image Database

Built for...

Clinicians & Educators

wishing to...

View Dermatology Images

Built for...

Clinicians

wishing to...

Contribute Images

Built for...

AI Researchers

wishing to...

Acquire Training Data

Built for...

AI Researchers

wishing to...

Take Part in ML Competitions

MicrosoftTeams-image.jpg

Engaging Stakeholder Communities

2024 Challenge

ISIC Machine Learning Challenge

In this competition, you'll develop image-based algorithms to identify histologically confirmed skin cancer cases with single-lesion crops from 3D total body photos (TBP). The image quality resembles close-up smartphone photos, which are regularly submitted for telehealth purposes. Your binary classification algorithm could be used in settings without access to specialized care and improve triage for early skin cancer detection.

The ISIC Archive

A large and expanding open-source public-access archive of skin images serves as a public resource for teaching, research, and the development and testing of diagnostic artificial intelligence algorithms

Browse tens of thousands of public images
Define an image collection and collect annotations
Access data through the command line
Contribute data from your own clinic

Serving the Clinical and Computer Vision Communities

The International Skin Imaging Collaboration (ISIC) is an academia and industry partnership designed to use digital skin imaging to help reduce skin cancer mortality.

 

ISIC works to achieve its goals through the development and promotion of standards for digital skin imaging, and through engaging the dermatology and computer vision communities toward improved diagnostics. 

MicrosoftTeams-image.jpg

Creating and Disseminating Skin Imaging Standards

Upcoming Workshop

The Alignment Problem in Medicine

In this special session, we will explore the alignment of AI systems with human values within image-based diagnostic medicine. To integrate complex ethical principles and embed human values into AI algorithm, it is required to understand ethical frames in healthcare and apply it in AI development. Transparency and biases in AI systems will also be important discussion topics in this session.

ids-logon_edited.png
20211122_DermoscopediaLogo.jpeg
SIIM_logo-600x315.png

Disseminating Standards

Compatibility in resolution, color processing, technical metadata, compression, and encryption

Josep Malvehy, MD
Hospital Clinic of Barcelona
Barcelona, Spain

Developing standards to ensure quality, privacy, and interoperability of dermatologic images.

Organizing workshops and ML challenges to engage with computer vision researchers.

Disseminating resources for educating the next generation of skin cancer experts.

Hosting competitions to engage the computer vision community to improve dermatologic diagnostic accuracy with the aid of AI.

  • Task - classification of lesions cropped from total-body photography images

    Featured Kaggle Competition

    June 26-September 6, 2024

    Released datasets​​

    • Training: 401,059
    • Testing: undisclosed

    Total prizes - $80,000

    Conference - MICCAI

  • Task - classification with clustered observations

    Melanoma vs benign

    Participation - 3,308 participant teams

    Released datasets​​

    • Training: 33,126
    • Testing: 10,982

    Total prizes - $30,000

    Conference - C-MIMI

  • Task - classification with out-of-distribution

    8 diagnoses + 1 OOD class

    Participation - 64 participant teams

    Released datasets​​

    • Training: 25,331
    • Testing: 8,238

    Total prizes - $7,000

    Conference - MICCAI

  • 3 Subtasks

    Lesion mask segmentation

    Attributes detection

    5 classes

    Diagnostic classification

    7 classes

    Participation - 113 participant teams

    Released datasets​​

    • Training: 10,015
    • Testing: 1,512

    Total prizes - $7,500

    Conference - MICCAI

  • 3 Subtasks

    Lesion mask segmentation

    Attributes detection

    4 classes

    Diagnostic classification

    3 classes

    Participation - 33 participant teams

    Released datasets​​

    • Training: 2,000
    • Testing: 600

    Conference - ISBI

  • Participation - 44 participant teams

    Released datasets​​

    • Training: 900
    • Testing: 379

    Conference - ISBI

    3 Subtasks

    Lesion mask segmentation

    Attribute detection

    2 morphologic features

    Diagnostic classification

    Melanoma vs bengin

  • The following tasks are open for live-submission scoring:

    • 2020: SIIM-ISIC Melanoma Classification with patient-clustered images (2 classes)

An overview of the ISIC Archive, voiceover by Veronica Rotemberg (AI Working Group Leader)

ISIC Working Groups have published dozens of seminal papers in high-impact journals.

ISIC has held numerous workshops at conferences including CVPR and MICCAI

4308215_business_conference_event_presentation_room_icon.png
MicrosoftTeams-image.jpg

Our Partners

MicrosoftTeams-image.jpg
8,427 Registered Users
81,722 Public Images
1,154,492 Total Images
21 Core Publications
1000+ Citations

Generously funded by The Shore Family Fund

The International Skin Imaging Collaboration​

Improving Skin Cancer Diagnosis by
 ● Promoting Standards in Skin Imaging

 ● Gathering and Sharing Dermatologic Images

 ● Engaging Clinicians & Computer Vision Researchers

bottom of page