Cilika TB — AI-Powered Tuberculosis Microscopy

Cilika TB is a professional iPad application designed for laboratory technicians and pathologists performing sputum smear microscopy for tuberculosis (TB) diagnosis.

AI-Assisted AFB Detection

Powered by on-device machine learning, Cilika TB automatically detects and counts Acid-Fast Bacilli (AFB) in real time through the iPad camera. Detected bacilli are highlighted with bounding boxes as the user scans microscope fields, helping reduce oversight and significantly improving examination efficiency. All AI processing is performed locally on the device and does not require an internet connection.

Advanced Camera Controls

Achieve consistent, high-quality image capture with precise control over exposure, white balance, and focus. Volume-button shutter support minimizes device movement during image acquisition, improving stability when capturing microscope fields of view (FOVs).

Patient and Accession Management

Manage patient accessions with complete traceability. Each accession record includes:

• Patient ID
• TB grading (Scanty, 1+, 2+, 3+)
• Pus cell count
• Laboratory observations
• Diagnostic comments

All information is stored in an organized and searchable format.

FOV Review and Manual Verification

Review every captured field-of-view image before finalizing results. Users can manually add, modify, or remove detected AFB annotations, ensuring accurate counts and maintaining expert oversight throughout the examination process.

Adjustable AI Confidence Threshold

Customize the AI detection confidence threshold to suit laboratory requirements. Higher thresholds prioritize precision, while lower thresholds surface additional candidate detections in challenging samples.

PDF Reporting and Data Export

Generate professional PDF reports directly within the application for individual accessions. Export patient and examination data in CSV or XLSX formats for record keeping, audits, reporting, and downstream analysis.

Secure and Offline-First Architecture

Patient data is stored locally on the device using SwiftData. User authentication is securely managed through Firebase, while all AI inference is executed entirely on-device. Patient data remains under local control and is not transmitted for AI processing.

Intended Use

Cilika TB is intended for use by trained laboratory personnel as a decision-support tool during sputum smear microscopy workflows. The application assists with AFB detection and documentation but is not intended to function as a standalone diagnostic device. Final clinical interpretation and diagnosis remain the responsibility of qualified healthcare professionals.