DERMOFIT:
A cognitive prosthesis to aid focal skin lesion diagnosis
The 10 most commonly observed skin lesions
 |
 |
 |
 |
 |
| Actinic |
Basal Cell |
Melanocytic |
Seborrhoeic |
Squamous Cell |
| Keratosis |
Carcinoma |
Nevus (mole) |
Keratosis |
Carcinoma |
| A premalignant lesion that usually requires treatment |
An invasive skin cancer that requires treatment (urgent/very soon) |
A normal finding in adults |
A benign skin tumour (treated for aesthetic reasons on occasions) |
An invasive skin cancer with a risk of distant spread (urgent) |
 |
 |
 |
 |
 |
| Intraepithelial |
Pyogenic |
Haemangioma |
Dermatofibroma |
Malignant |
| Carcinoma |
Granuloma |
|
|
Melanoma |
| A premalignant lesion that may develop into a squamous cell carcinoma |
A benign lesion, that often requires treatment because of bleeding |
A benign collection of blood vessels |
A benign dermal tumour |
The most serious type of skin cancer (urgent) |
BRIEF PROJECT OVERVIEW
DERMOFIT is a Wellcome Foundation funded research project.
The project goal is to develop an interactive skin cancer image
database indexing tool, where users compare 'live' skin lesions/spots
to images selected from a database. Based on user feedback on the
goodness of the selections, the database will searched for better
matches.
The project has both science and social goals.
The science goal is to make contributions to the diagnosis
of non-melanoma cancers (which are not widely studied by other teams,
as contrasted with melanoma, which is widely studied).
The social goal is to develop a web-based tool that can be
used by non-specialist medical people (e.g. GPs, clinic nurses)
for pre-assessment
of spots, or even potentially by the general public as a tool
to encourage visits to their doctor if they are worried about a
spot.
Previous exploratory projects (Banks, Stavrakakis) developed
database extraction algorithms and an interactive interface
that looks like:
This project will build on those studies, aiming to:
- develop algorithms for classification of these types of lesions,
- build a publically usable web-based self-assessment tool and
- collect a database of usable skin samples for other skin lesion
researchers.
The project starts on August 1, 2008 and continues until July 31, 2011.
PROJECT PLANS
The key components of this project are:
- Clinical collection of many colour and 3D shape images from
a wide variety of skin lesions.
- Mounting these images for public use (on the web).
- Development of a variety of colour, shape and texture image properties.
- Machine learning algorithms to discover the most usable features, plus
to discover the latent patterns of likeness based on user interaction.
- Java interface extension for web use and new capabilities.
- Clinical and web-based diagnostic assessment.
- Usability assessment.
RESEARCH RESULTS
You can find some:
- Publically downloadable datasets.
- Project publications
- Our skin cancer
research has led to the DERMOFIT
skin cancer training and diagnostic assistant iPad app in partnership with
Simedics.
RESEARCH POSITIONS AVAILABLE
At the moment, there are no paid positions. Visiting researchers are welcome.
PRINCIPAL INVESTIGATORS
Robert Fisher
School of Informatics
Univ. of Edinburgh
1.26 Informatics Forum
10 Crichton St
Edinburgh EH8 9AB, United Kingdom
Phone: (44)-131-651-3441 (direct line)
Phone: (44)-131-651-3443 (secretary)
Fax: (44)-131-650-6899
Email: 
http://www.inf.ed.ac.uk/people/staff/Robert_Fisher.html
|
Jonathan Rees
Dermatology
The University of Edinburgh
Lauriston Building
Lauriston Place
Edinburgh EH3 9HA
Tel: 44-(0) 131 536 2041
Fax: 44-(0) 131 229 8769
Email: jonathan.rees@ed.ac.uk
http://www.derm.med.ed.ac.uk/05_staff/rees.htm
|