Tweris - AI-Powered CTG Fetal Monitoring

During labour, clinicians manually interpret cardiotocography (CTG) readings to assess fetal wellbeing. This process is slow, subjective, and inconsistent - different clinicians can reach different conclusions from the same trace. Tweris needed to convert raw CTG signals into standardized, clinically validated assessments that any clinician can trust, delivered in seconds, not minutes.

Signal Type: Cardiotocography (CTG) - fetal heart rate and uterine contraction patterns. Complex, high-volume, noisy signals with frequent data dropouts.

1. Signal Extraction from CTG Images Computer vision algorithms extract and digitize fetal heart rate and contraction waveforms directly from uploaded CTG images. The system handles varying image quality, different CTG machine output formats, and incomplete traces.

2. Baseline Modeling and Pattern Detection AI models detect and quantify FHR baseline, beat-to-beat variability, accelerations, and decelerations. The system classifies deceleration types (early, late, variable) and assesses clinical significance against the contraction pattern.

3. Contextual Clinical Analysis The AI cross-references signal patterns with patient-specific context: gestational age, maternal age, and clinical history. No signal is analyzed in isolation — every assessment accounts for the full clinical picture.

4. Structured Diagnosis and Documentation Model results are translated into a structured fetal wellbeing diagnosis with evidence-based clinical recommendations. Results are exportable as PDF for documentation and archiving. Consistent, reproducible assessments regardless of which clinician reviews the output.

The Interface: Conversational AI for Clinicians We built the entire system around an intuitive chat interface. Clinicians upload a CTG image, input patient context, and receive a comprehensive fetal wellbeing assessment within seconds — no specialized technical knowledge required. Chat history and results export as PDF for clinical records.

To address the challenges at hand, Rubix Code and Stealth Startup jointly embarked on an innovative project employing Physiological CTG analysis. The resulting application features a user-friendly interface with a chatbot, allowing easy upload of CTG images and relevant information. Using advanced AI algorithms, the application conducts a comprehensive analysis, estimating baselines and considering contextual factors such as the week of pregnancy and the age of the mother.

• 40% faster CTG analysis compared to manual interpretation
• Consistent diagnostic outputs eliminating inter-observer variability
• Automated processing reduces human error during high-pressure labour
• Evidence-based recommendations are paired with every assessment
• Exportable clinical documentation from every interaction

Client Quote

"The team has been developing our innovative AI application in a very clever, relevant, time-saving, and cost-effective way. They take up the technological challenges with enthusiasm." — Luka Velemir, CEO and Founder, Tweris