Emphysema classification using convolutional neural networks

Abstract

There has been paid more and more attention in diagnosing emphysema using High-resolution Computed Tomography. This may lead to improve both understanding and computer-aided diagnosis. We propose a novel classification framework using convolutional neural network(CNN). This model automatically extracts features from the raw image and generates classification. Experiments have been conducted on the database from clinical. Results a recognition rate of 92.54% for classification two kinds of emphysema with normal. The designed convolutional neural networks can get better results for classifying one kind of emphysema with normal.