Can Edge Map be Sufficient for Deep Learning Models to Understand Chest Radiographs?

Presentation Type

Poster

Student

Yes

Track

Health Care Application

Abstract

In computer vision, edge maps can help understand line-rich objects, such as buildings and cars when image data are considered. The edge map does not include color/texture properties in it. Organizing their relative positioning of these lines/curves is not trivial, and therefore the use of deep learning could possibly avoid missing information. In this work, using deep learning models, we present the use of an edge map for understanding abnormalities (such as Tuberculosis and Pneumonia) in chest X-rays. Our results will be followed by the discussion, where we state the primary motivation behind the use of the edge maps, not the textures.

Start Date

2-11-2020 1:00 PM

This document is currently not available here.

Share

COinS
 
Feb 11th, 1:00 PM

Can Edge Map be Sufficient for Deep Learning Models to Understand Chest Radiographs?

Volstorff A

In computer vision, edge maps can help understand line-rich objects, such as buildings and cars when image data are considered. The edge map does not include color/texture properties in it. Organizing their relative positioning of these lines/curves is not trivial, and therefore the use of deep learning could possibly avoid missing information. In this work, using deep learning models, we present the use of an edge map for understanding abnormalities (such as Tuberculosis and Pneumonia) in chest X-rays. Our results will be followed by the discussion, where we state the primary motivation behind the use of the edge maps, not the textures.