MaChestLocal

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

Localization of disease with limited supervision in chest radiographs

Supervision: Prof. Dr. Nassir Navab, Dr. Seong Tae Kim, Ashkan Khakzar

Abstract

Chest radiographs are known as the most widely used type of scan in the world. Developing a computer-aided diagnosis can help the radiologists by decreasing the diagnostic error rate and reducing the reading time, which eventually increases the overall performance of the workflow. Recently, a large research effort has been devoted to developing an automatic diagnosis method with publically available chest radiography datasets. It is important to accurately diagnose and localize the disease at the same time. The performance of deep learning highly depends on the label in the training dataset. However, it is a very time-consuming and expensive task to get an annotation from medical experts for the location of the disease in chest radiographs. In this project, a novel way with limited supervision to localize the disease in chest radiographs will be investigated.

Requirements:

  • Good understanding of statistics and machine learning methods.
  • Very good programming skills in Python & TensorFlow? / PyTorch?

Location:

  • Garching


ProjectForm
Title: Localization of disease with limited supervision in chest radiographs
Abstract: Chest radiographs are known as the most widely used type of scan in the world. Developing a computer-aided diagnosis can help the radiologists by decreasing the diagnostic error rate and reducing the reading time, which eventually increases the overall performance of the workflow. Recently, a large research effort has been devoted to developing an automatic diagnosis method with publically available chest radiography datasets. It is important to accurately diagnose and localize the disease at the same time. The performance of deep learning highly depends on the label in the training dataset. However, it is a very time-consuming and expensive task to get an annotation from medical experts for the location of the disease in chest radiographs. In this project, a novel way with limited supervision to localize the disease in chest radiographs will be investigated.
Student: Miraç Sanisoğlu
Director: Prof. Dr. Nassir Navab
Supervisor: Dr. Seong Tae Kim, Ashkan Khakzar
Type: Project
Area: Machine Learning, Medical Imaging
Status: running
Start:  
Finish:  
Thesis (optional):  
Picture:  


Edit | Attach | Refresh | Diffs | More | Revision r1.2 - 22 Oct 2019 - 20:08 - SeongTaeKim