4@DBE'2020 Mobile application for imaging and classification of skin lesions

project type: classic

edition: 2020

number of students in project 3 - 4

manager: Martyna Giziewicz

The purpose of the group project is to develop, implement and evaluate a mobile application (Android) for the assessment of skin lesions, in particular hyperpigmentation, for the diagnosis of skin melanoma. For the classification of skin lesions, artificial intelligence and machine learning methods should be used. Learning collections will be obtained from the archives of the Medical University of Gdańsk and other medical image databases.

The ready application is to be a tool for the doctor. It is assumed that skin lesions and the progress of these changes will be assessed during subsequent patient visits to a specialist clinic. The basic set of parameters analyzed: size of the change, shoreline features, asymmetry, convexity, color).

The task will also require the development of a platform for storage, sharing and description of tests.

It is assumed to use available libraries and SDK packages (Android, OpenCV, TensorFlow, etc.).

Tasks to perform:

1.     Analysis of the state of knowledge in the field of image analysis in dermatology,


2.     Analysis of the state of knowledge in the field of artificial intelligence and machine learning,


3.     Development of a database schema design for the collection of data on test results,


4.     Development of a project of an internet system for collecting and sharing data,


5.     NN design and learning,


6.     Development of a mobile application project for Android,


7.     Implementation of projects,


8.     Carrying out tests on clinical data,


9.     Preparation of documentation, including user and administrator documentation.