Tuberculosis is an infectious disease that leads to the death of millions of people across the world. The mortality rate from this disease is high in patients suffering from immunocompromised disorders. early diagnosis can save lives and avoid further complications. However, the diagnosis of TB is a very complex task. The standard diagnostic tests still rely on traditional procedures developed in the 20th century. These procedures are slow and expensive. Therefore, this chapter presents an automatic approach for the diagnosis of TB from posteroanterior chest X-rays. This is a two-step approach, in which in the first step the lung regions are segmented from the chest X-rays using the graph cut method, and then in the second step the transfer learning of VGG16 combined with bidirectional LSTM is used for extracting high-level discriminative features from the segmented lung regions and then classification is performed using a fully connected layer. The proposed model is evaluated using data from two publicly available databases, namely the Montgomery County set and the Schezien set. The proposed model achieved accuracy and sensitivity of 97.76%, 97.01% and 96.42%, 94.11% on the Schezien and Montgomery County datasets, respectively. This model enhanced the diagnostic accuracy of TB by 0.7% and 11.68% on the Schezien and Montgomery County datasets, respectively.

Chapter Contents:

• Abstract
• 3.1 Introduction
• 3.2 Related works
• 3.3 Methods
• 3.3.1 Data preprocessing
• 3.3.2 Proposed methodology
• 3.4 Results and discussion
• 3.4.1 Databases
• 3.4.2 Performance metrics
• 3.4.3 Results reported
• 3.4.4 Comparison
• 3.5 Conclusion
• References

Inspec keywords: diseases; image segmentation; feature extraction; image classification; lung; medical image processing; diagnostic radiography; learning (artificial intelligence)

Other keywords: VGG16; infectious disease; class dependency-based; transfer learning; Montgomery County set; immunocompromised disorders; high-level discriminative features; mortality rate; standard diagnostic tests; segmented lung regions; Bi-LSTM; tuberculosis; bidirectional LSTM; posteroanterior chest X-rays; graph cut method

Subjects: Machine learning (artificial intelligence); Patient diagnostic methods and instrumentation; Computer vision and image processing techniques; Image recognition; Biology and medical computing; X-ray techniques: radiography and computed tomography (biomedical imaging/measurement); X-rays and particle beams (medical uses)