Tuberculosis Detection Using Chest X-Ray Image Classification by Deep Learning

Tuberculosis (TB) is a deadly and widespread lung disease that is often not easily detectable in the early stages. Thanks to the availability of high-resolution chest X-rays, deep learning (DL) is now able to help with the successful detection of this malignant disease, along with other possible applications in the health sector. In this manuscript, a new deep-learning model for TB detection is proposed using chest X-ray image classification. To achieve this, a mixture of two popular pre-trained deep learning CNNs has been employed (VGG16 and VGG19) utilizing the ImageNet dataset, in addition to the block attention module to obtain spatial data. This method has been proven to be valid through experiments on four popular Datasets; NLM dataset, Belarus dataset, NIAID TB dataset, and RSNA-CXR dataset. The evaluation showed results in achieving an excellent accuracy of 0.9966 and 0.9978 for both training and validation sets respectively.

• tuberculosis detection
• chest X-ray classification
• transfer learning
• VGG16
• VGG19
• pre-trained convolutional neural networks
• lung disease diagnosis
• medical image analysis

• 3 — Bonne santé et bien-être

• COVID-19 diagnosis using AI
• AI in cancer detection
• Tuberculosis Research and Epidemiology
• tuberculosis detection
• chest X-ray classification
• transfer learning
• VGG16
• VGG19
• pre-trained convolutional neural networks
• lung disease diagnosis
• medical image analysis
• Radiology, Nuclear Medicine and Imaging
• Artificial Intelligence
• Infectious Diseases
• Medicine
• Computer Science