Diagnosis of secondary pulmonary tuberculosis by an eight-layer improved convolutional neural network with stochastic po
- PDF / 3,956,066 Bytes
- 18 Pages / 595.276 x 790.866 pts Page_size
- 36 Downloads / 169 Views
ORIGINAL RESEARCH
Diagnosis of secondary pulmonary tuberculosis by an eight‑layer improved convolutional neural network with stochastic pooling and hyperparameter optimization Yu‑Dong Zhang1,2 · Deepak Ranjan Nayak3 · Xin Zhang4 · Shui‑Hua Wang5,6 Received: 7 July 2020 / Accepted: 10 October 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract To more efficiently diagnose secondary pulmonary tuberculosis, we build an improved convolutional neural network (ICNN) based on recent deep learning technologies. First, a 12-way data augmentation (DA-12) was proposed to increase size of training set. Second, stochastic pooling was introduced to replace the standard average pooling and max pooling. Third, batch normalization and dropout techniques were included and associated with conv layers and fully-connected layers, respectively. Fourth, a dynamic learning rate was employed to replace traditional fixed learning rate. Fifth, hyperparameter optimization was used to optimize the number of layers within proposed network. Our eight-layer ICNN demonstrated excellent results on the test set, yielding a sensitivity of 94.19%, a specificity of 93.72%, and an accuracy of 93.95%. Our ICNN provides better performances than other four state-of-the-art algorithms. It can help radiologists to make more accurate diagnosis on secondary pulmonary tuberculosis. Keywords Secondary pulmonary tuberculosis · Deep learning · Convolutional neural network · Stochastic pooling · Dynamic learning rate · Hyper-parameter optimization Abbreviations (D)/(F)LR (Dynamic)/(fixed) learning rate HC Healthy control TB Tuberculosis PT Pulmonary TB EPT Extrapulmonary TB PPT Primary pulmonary tuberculosis Yu-Dong Zhang and Deepak Ranjan Nayak contributed equally to this paper. * Xin Zhang [email protected] * Shui‑Hua Wang [email protected] Yu‑Dong Zhang [email protected] Deepak Ranjan Nayak [email protected] 1
SPT Second pulmonary tuberculosis CXR Chest X-ray IS Immune system SPT Secondary pulmonary tuberculosis HS Histogram stretching CR Compression ratio SSDP Small-size dataset problem LoG Lack of generalization DA(12) Data augmentation (12-way) GC Gamma correction RT Random translation NI Noise injection (H)(V)ST (Horizontal) (vertical) shear transform 3
Department of Computer Science and Engineering, Malaviya National Institute of Technology, Jaipur 302017, India
4
Department of Medical Imaging, The Fourth People’s Hospital of Huai’an, Jiangsu Province, Huai’an 223002, China
School of Informatics, University of Leicester, Leicester LE1 7RH, UK
5
School of Architecture Building and Civil Engineering, Loughborough University, Loughborough LE11 3TU, UK
Department of Information Systems, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah 21589, Saudi Arabia
6
Department of Cardiovascular Sciences, University of Leicester, Leicester LE1 7RH, UK
2
13
Vol.:(0123456789)
CL Conv layer PL Pooling layer FCL Fully connected layer ReLU Rectified linear
Data Loading...
