This infographic showcases the user interface of the TB-DRC-DSS, a web-based decision support system for tuberculosis (TB) detection and drug resistance classification, displayed across four mobile device frames. Each screen follows a consistent layout: a header titled 'TB DETECTION AND DRUG RESISTANT-TB CLASSIFICATION SYSTEM', an image upload button, a central posterior-anterior (PA) chest X-ray (CXR), and a diagnostic results section at the bottom. The visual demonstrates four distinct clinical categories: Drug-Susceptible TB (DS-TB), Drug-Resistant TB (DR-TB), Multidrug-Resistant TB (MDR-TB), and Extensively Drug-Resistant TB (XDR-TB). Each X-ray displays varying pulmonary pathologies, including consolidations, cavitary lesions, and diffuse opacities. Below each radiographic image, the system provides a specific 'CLASS' label and a 'RECOMMENDED REGIMEN' according to established treatment protocols (e.g., 2HRZE/4HR for DS-TB or Bedaquiline-containing regimens for MDR-TB). The tool is designed for pulmonologists and TB specialists to integrate diagnostic imaging with pharmaceutical decision-making at the point of care.

This infographic depicts a mobile-based diagnostic support system for tuberculosis (TB) management, presenting three separate smartphone interface frames. Each frame features a posterior-anterior (PA) chest X-ray (CXR) with AI-generated classifications and treatment regimens. The first image shows significant opacification and consolidation in the upper lung fields, classified as Drug-Resistant TB (DR-TB), with complex multidrug regimens listed. The second image displays relatively clear lung fields with normal vascular markings, classified as Drug-Susceptible TB (DS-TB), suggesting early-stage disease or a normal baseline, paired with a standard '2HRZE/4HR' regimen. The third image illustrates mild interstitial markings and apical opacities, classified as Multi-Drug Resistant TB (MDR-TB), recommending a shorter all-oral bedaquiline-containing regimen. This visual demonstrates the clinical application of deep learning algorithms in identifying TB drug-response phenotypes (DS, DR, MDR) and providing automated therapeutic guidance based on WHO protocols for infectious disease specialists and healthcare providers.

This infographic presents the mobile user interface (UI) of the TB-DRD-CXR application, an artificial intelligence tool designed for diagnosing drug-resistant tuberculosis (TB) using chest X-rays (CXR). The UI is displayed across five screens. The top row includes the 'Home page' with introductory text, followed by 'Sign up' and 'Log in' registration modules. The bottom row demonstrates the diagnostic functionalities: 'Still image analysis' and 'Fast image analysis'. The 'Still image analysis' screen allows for local file uploads, displaying the identified classification (e.g., DR-TB) and suggesting associated clinical treatment regimens. The 'Fast image analysis' screen features a live-camera interface that performs real-time classification. It provides a probability distribution for five distinct categories: drug-susceptible TB (DS-TB), drug-resistant TB (DR-TB), multidrug-resistant TB (MDR-TB), extensively drug-resistant TB (XDR-TB), and pre-extensively drug-resistant TB (Pre-XDR). These probabilities are visually represented using color-coded progress bars to assist medical professionals in rapid triage and diagnosis.

A comparison chart demonstrating a live image analysis interface for tuberculosis (TB) drug resistance diagnosis from chest X-rays (CXR). The graphic is divided into four chronological columns: 0.00 seconds, 5.00 seconds, 10.00 seconds, and shutter pressing time at 15 seconds. Each column displays a cropped diagnostic image focusing on the bilateral mid-to-lower lung fields and mediastinal borders. Below each image are AI-generated probability bars for five clinical classifications: Pre-XDR (pre-extensively drug-resistant TB), DR-TB (drug-resistant TB), DS-TB (drug-susceptible TB), XDR-TB (extensively drug-resistant TB), and MDR-TB (multidrug-resistant TB). The visual illustrates the real-time evolution of the deep learning model's diagnostic confidence. For instance, at 5.00 seconds, XDR-TB shows a 100% probability, while at the final shutter time (15 seconds), DS-TB becomes the most likely prediction at 66%. This infographic highlights the application of 'fast image analysis' in clinical informatics and infectious disease screening.