A New Machine Learning System That Automatically Classifies the Shapes of Red Blood Cells

Using a computational approach referred to as deep learning, researchers have developed a new system that identifies the shapes of red blood cells. Published in PLOS Computational Biology, these findings could potentially assist doctors to monitor individuals with sickle cell disease.
People with sickle cell disease produce red blood cells that are abnormally shaped. These cells can build up and blood vessels in the body, causing pain and at times death. Also referred by some people as sickle-cell anemia, sickle cell disease is named after crescent-like (sickle-shaped) red blood cell. However, it results in several other shapes such as elongated or oval red blood cells. Although the shape of red blood cells in a particular patient can determine the severity of their disease, it is difficult to manually identify these shapes.
Mengjia Xu of Northeastern University, China, and his colleagues have developed a computational framework that uses a machine-learning tool called a deep convolutional neural network (CNN). The framework automates the process of classifying the shapes of the red blood cells.
The new framework employs three steps to identify the red blood cells shapes in microscopic images of blood. It distinguishes the cells from each other and the background of each image. Then for each cell identified, it zooms in or out until the cells images are uniform in size. Lastly, it uses a deep convolutional neural network to classify the cells by shape.


답글 남기기

아래 항목을 채우거나 오른쪽 아이콘 중 하나를 클릭하여 로그 인 하세요:

WordPress.com 로고

WordPress.com의 계정을 사용하여 댓글을 남깁니다. 로그아웃 /  변경 )

Google+ photo

Google+의 계정을 사용하여 댓글을 남깁니다. 로그아웃 /  변경 )

Twitter 사진

Twitter의 계정을 사용하여 댓글을 남깁니다. 로그아웃 /  변경 )

Facebook 사진

Facebook의 계정을 사용하여 댓글을 남깁니다. 로그아웃 /  변경 )

%s에 연결하는 중