Predicting Tuberculosis Related Lung Deformities from CT Scan Images Using 3D CNN

Anup Pattnaik, Sarthak Kanodia, Rahul Chowdhury, Smita Mohanty
2019 Conference and Labs of the Evaluation Forum  
CT scan of lung has become an invaluable tool in the diagnosis of tuberculosis. However, analysis of 3-D image data is time consuming and relies heavily on trained expertise. As an attempt to automate this approach without compromising accuracy advanced AI algorithms have been explored to draw clinically actionable hypothesis. The approach comprises of detailed image processing, followed by feature extraction using tensor flow and 3-D CNN to further augment the metadata with the features
more » ... ed from the image data and finally perform 6 class binary classification using random forest. On the test dataset the method resulted in an overall mean AUC of 0.6.
dblp:conf/clef/PattnaikKCM19 fatcat:bfynuo4dxvhpfbfiuktgolu2ba