logo

Menu

Tracks

SpecialSession: AI-Driven Real-Time Monitoring and Anomaly Detection in Biomedical Data Streams

AI-Driven Real-Time Monitoring and Anomaly Detection in Biomedical Data Streams

Ali ŞenolAli Şenol
Garima AgrawalGarima Agrawal
Tarık TalanTarık Talan
Huan LiuHuan Liu

The proliferation of wearable sensors, continuous monitoring devices, and Internet of Medical Things (IoMT) technologies has created unprecedented opportunities and challenges in healthcare. The global IoMT market, valued at approximately $65 billion in 2025, is projected to reach $155 billion by 2030, reflecting the rapid adoption of real-time health monitoring systems. These technologies generate massive volumes of high-frequency biomedical data streams that require real-time analysis for critical decision-making.

This special session addresses the computational and algorithmic challenges of processing, analyzing, and extracting actionable insights from streaming biomedical data. The shift from hospital-centric to patient-centric care, accelerated by telemedicine and wearable technologies, demands new computational paradigms that can process data on-the-fly while maintaining clinical accuracy and reliability.

This session bridges the gap between traditional bioinformatics and the emerging field of streaming health analytics. While genomic analysis deals with large but static datasets, IoMT systems produce endless streams of time-series data that must be analyzed with minimal latency. The unique challenges include:

  • Temporal constraints requiring decisions in milliseconds to minutes
  • Resource limitations on edge devices with limited computational power
  • Concept drift as patient baselines change over time
  • Balance between false negatives and alarm fatigue (ICU false alarm rates: 80-99%)

Topics of interest include, but are not limited to:

Real-time processing architectures for physiological signals (EEG, ECG, EMG, PPG)Streaming algorithms for high-dimensional biological dataEdge computing and fog computing solutions for biomedical applicationsReal-time anomaly detection in physiological time seriesEarly warning systems for critical health eventsMulti-modal sensor fusion for anomaly identificationHandling non-stationarity in patient monitoring dataAdaptive algorithms for evolving health conditionsPersonalized baseline detection and transfer learning for patient-specific modelsOnline clustering methods for continuous health monitoringIncremental classification techniques for real-time diagnosisFeature selection and dimensionality reduction in streaming contextsInterpretable AI models for clinical decision supportIntegration of domain knowledge with data-driven approachesAlarm management and reducing false positives in ICU settingsRemote patient monitoring, wearable health devices, and telemedicineICU monitoring, early sepsis detection, and cardiac arrhythmia detectionSleep disorder identification and seizure prediction

Contacts