Unsupervised Machine Learning for Cancer Detection Using Multimedia Data

Master

In this project, we aim to design and develop a system for analysing video from a camera pill. The pill is swallowed and records video of the digestive system - the goal is to be able to automatically detect cancer in the colon. The master student will explore unsupvervised methods in machine learning to detect cancer and possible other diseases.

Goal

The goal is to develop a system that can make sense in the large amount of data that we have. At the moment we are also working on a method to get annotations for this data. An unsupervised machine learning method can be very helpful for the annotation part too.

Learning outcome

State of the art in image and video processing and machine learning. Apart from that the student will also get familiar with the medical background.

Qualifications

  • Signal processing
  • Good programming skills

Supervisors

  • Pål Halvorsen
  • Michael Riegler

References

PillCam - A Camera Pill System for Automatic Screening of the Digestive System