Application of Deep Learning on CXR Data to Signal Cases of Pneumonia
Blaku, Flokart; Saidnassim, Nurbek (2018)
Hämeen ammattikorkeakoulu
2018
All rights reserved
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2018112718402
https://urn.fi/URN:NBN:fi:amk-2018112718402
Tiivistelmä
The recent wave of developments in artificial intelligence and data science inspired us to work on this thesis topic that challenged us and developed our skills.
The aim of this thesis was to implement an algorithm that would detect the presence of pneumonia in patients. This project was proposed by The Radiological Society of North America as an open source challenge on the renown data science platform Kaggle. The content of the thesis is mainly focused on the implementation, theory, methodology and results. Additionally, some obstacles that came up during the process are included.
Our efforts resulted in a system with an accuracy of 82% and the designed detector ranked us on the top 15% solutions in Kaggle.
The aim of this thesis was to implement an algorithm that would detect the presence of pneumonia in patients. This project was proposed by The Radiological Society of North America as an open source challenge on the renown data science platform Kaggle. The content of the thesis is mainly focused on the implementation, theory, methodology and results. Additionally, some obstacles that came up during the process are included.
Our efforts resulted in a system with an accuracy of 82% and the designed detector ranked us on the top 15% solutions in Kaggle.