Pneumonia Detection with Deep Convolutional Architecture

摘要

Pneumonia is a respiratory disease caused by an infection in the air sacs of the lungs. Patient with pneumonia will experience inflammation of the alveoli, accompanied by the presence of fluids in the air sacs. Using the intensity from thorax x-ray images, a radiologist can diagnose whether pneumonia exist or not. Computer-aided detection (CAD) can enhance the radiologist diagnostic capabilities by giving radiologist a second opinion. CAD system can be developed by several techniques including deep convolutional architecture. This paper aims to know the performance of two widely known deep convolutional architecture such as residual network and mask-RCNN in classifying and detecting pneumonia. In addition, the results will be compared and evaluated.