A melanoma diagnosis and prognosis framework with human readable explanation

Melanoma is the most lethal skin cancer, accounting for 2% of all skin cancer types, yet approximately 75% of skin cancer deaths. It often evolves from clear skin or existing moles, making it difficult to diagnose at early stage. Besides, the treatment of melanoma is a complex decision making process, which is affected by a large number of internal and external factors, e.g. disease location, staging, etc. Our objective is to utilize the medical data collected by CMRN to design an electronic tool to save valuable time of clinicians in routine pathology assessment and ultimately assist evidence based decision making. The system will help to identify high risk pathological observations. It combines state-of-the-art pattern recognition algorithms with natural language processing (NLP), which generates human readable explanation. Moreover, our system can generate easy-understandable analysis results to patients, which enables patients to better understand their own conditions. TO BE CONT’D

Faculty Supervisor:

Scott Ernst

Student:

Xue Teng

Partner:

Pulse InfoFrame Inc.

Discipline:

Dentistry

Sector:

Medical devices

University:

Program: