Nearshore Application Development Services (Poland)

Case Study: Healthcare / Artificial Intelligence for Medical Image Classification

FIS SST nearshoring team looking into the research of using computer based classification of x-rays. Use of machine learning to potentially detect lung cancer in patients

X-ray of lungs

Introduction

FIS SST worked on a research project, looking to establish if machine learning could be used to provide a reliable method of classification between healthy and unhealthy x-ray images of patient lungs.

Two of the core technologies FIS SST employed for the machine learning project included:

Tensorflow - an end-to-end platform for training machine learning models

Keras - a deep learning framework technology.

Challenges

  • Data selection and x-ray images preparation for building the machine learning models
  • Access to reference data for health individuals
  • Lack of medical expert knowledge on the assessment of x-rays
  • Various file formats for the X-ray images