Detecting Skin Cancer with Low cost devices and AI technology

Skin  Diagnostic AI is an open source projects created to develop application for medical specialists that will help identifying potential skin cancer, and collect dermatoscopic images for medical community. It will also hep to patients to store their images in easy and accessible way on personal devices.

Learn How to run interactive session on renku

Current Status

Skin Daignostic AI is on early stage development, The Feasibility study, was completed, where I tested over 5000 different models and pipelines.

  • Apr. 2021 Deployment on Renku.io
  • Mar. 2021 AI Readiness Asessment study
  • Jan. 2021  End of Feasibility Study
  • Nov. 2021 PyClass adoption
  • Oct. 2021 Datasets Review

Key Features

  • Open Access (GPL license, other lince types are available on request) 
  • Basic Image analysis on mobile platform
  • Version for Patiens & Medical Specialists 

Feasibility Study

The goal is to quickly evaluate the dataset with large number of ml and deep learning models

  • to evaluate models with different diagnostic purposes
  • to test, of the shelf AI solutions
  • to explore different strategies for data preparation, treatment and feature extraction,
  • to develop baseline for further analyses,

Technology

SkinDiagnosticAI was implemented in Python using open source tools

Skin Diagnostic AI & PyClass
development timeline

Our Team

Pawel Rosikewicz picture

Pawel Rosikiewicz

Product Owner
& Lead data Scientists

Oksana Riba Grognuz

Senior Data Science Eng
Expert on Colabortatice platforms

Onur Yürüten

Data Science Manager
Business Intelligence

Supported by

Acknolwdments

  • Antonio Salvuci PhD. IT support
  • Marek Radwanski, Product development
  • Marta Rosikiewicz PhD. Medical QMS
  • Ieva Vaišnoraitė-Navikienė, AI solutions Implementation