Texture based interstitial lung disease detection using convolutional neural network
Pratiksha Hattikatti · 2017
Large range of lung texture patterns of disease can be observed in CT scan images. These images are the intermixed of various patterns and hence it becomes very difficult for Radiologist to differentiate between them and diagnose the disease. One way of solving this issue is use of Convolutional neural networks (CNN). CNN is generally used for pattern classification and image recognition systems. They have achieved less error on the database, image classification using CNN was surprisingly fast. Interstitial lung disease is a term which includes different types of lung disease. Interstitial lung diseases affect the interstitium i.e. the part of the lung's anatomic structure. Lung tissue characterization is essential parts of a computer aided diagnosis (CAD) system for detection of interstitial lung diseases (ILDs). Thus using CNN, interstitial lung disease detection gives accurate result. The proposed system is consists of CNN having 7 layers with Local binary pattern (LBP) as feature extractor. The execution of classification exhibited the capability of CNNs in analyzing lung patterns. The CT Scan Images used in this study are officially verified by the certified Radiologist.