Speakers
Description
We propose an edge-oriented clinical triage framework designed to streamline real-time risk stratification by fusing multimodal physiological data, specifically electrocardiogram (ECG) signals and chest X-ray inputs, into a unified severity verdict per patient. At its core, the system utilizes a PyTorch-based 1D Convolutional Neural Network (1D-CNN) trained on the MIT-BIH Arrhythmia dataset to classify 187-sample heartbeat windows into five distinct cardiac rhythm categories within a dedicated inference module. Data ingestion and inter-process communication are orchestrated through a NATS publish/subscribe messaging pipeline that emulates multi-sensor data streams, pairing incoming ECG and radiograph messages by patient identifier. Upon synchronization, a subscriber node executes the 1D-CNN model and evaluates composite telemetry via a programmatic function to yield an actionable diagnostic verdict categorized into Normal, Moderate, High, or Critical risk tiers. To ensure operational resiliency in decoupled or constrained network environments, the messaging layer incorporates an automated fallback mechanism for offline simulation. System persistence is managed via prana_database.py, which writes all classified verdicts and structured telemetry to an SQLite database alongside formatted, human-readable text reports. Real-time clinical visualization and monitoring are delivered via an auto-refreshing Streamlit and Plotly dashboard, presenting aggregate system metrics, severity distribution graphs, confidence histograms, prioritized critical-case tables, and searchable diagnostic records. Formulated around a modular “Raspberry Pi → NATS → Hailo AI → SQLite → Dashboard” edge computing architecture, the entire software framework is engineered such that migrating from laptop-based simulation to physical hardware deployment requires modifications strictly confined to the sensor-ingestion interface.
Session author's bio
Arghyajyoti Ghosh Mr. Arghyajyoti Ghosh is a 4th year undergraduate student of the dept of Electronics and Electrical Communication Engineering enrolled in its B. Tech. course. He hails from Durgapur, West Bengal.
Dhritish Mondal Mr. Dhritish Mondal is a 4th year undergraduate student of the dept of Electronics and Electrical Communication Engineering enrolled in its Dual degree course. He hails from Kolkata, West Bengal.
Priyam Chakraborty Priyam Chakraborty is an Assistant Professor in the Department of Aerospace Engineering at the Indian Institute of Technology Kharagpur. He holds B. Tech., M. Tech. and PhD, all three from IIT Kharagpur. He committed to the use of computational tools as a lead data scientist in a startup ecosystem, followed by post-doctoral fellowship at the University of Waterloo and the Indian Institute of Science. He is currently working on smart navigation of active matter in the department of Artificial Intelligence at IIT Kharagpur. His fascination with collective intelligence, initially sparked by the efficiency of bird flocks, has turned into a belief in the power of biomimetic design that invokes automation and machine learning to unlock hidden patterns within complex datasets. By exploring, analyzing, and potentially challenging existing assumptions, Priyam aims to create affordable intelligent systems that can automatically analyze video data, extract meaningful insights about human activities, and potentially be applied in areas like security monitoring, human-computer interaction, and medical diagnostics.
Suman Chakraborty Prof. Suman Chakraborty of IIT Kharagpur is a globally recognized expert in microfluidics and biomedical engineering, known for translating deep science into affordable healthcare solutions for underserved populations. A Sir J.C. Bose National Fellow and recipient of the 2026 TWAS Award (UNESCO), he has pioneered technologies like paper-and-pencil microfluidics and the COVIRAP molecular diagnostic platform. His work bridges fundamental fluid mechanics with low-cost medical diagnostics, including cancer screening tools and blood tests deployable in rural settings. With over 525 research publications, 25+ patents, and 50 Ph.D. graduates, his contributions span both fundamental science and impactful innovation. He leads a National CRTDH to promote indigenous medical device manufacturing and rural entrepreneurship. Prof. Chakraborty is a Shanti Swarup Bhatnagar Awardee, Infosys Prize winner, Fellow of top global academies, and a driving force in democratizing healthcare technologies by empowering local communities with science-driven solutions.
Any other info we should know?
Our proposed edge-oriented clinical triage framework intrinsically aligns with the conference’s dual focus on open-source computing and socioeconomic impact by democratizing advanced healthcare diagnostics. The architecture relies exclusively on accessible, community-driven technologies. By leveraging PyTorch for neural network modeling, NATS for robust telemetry messaging, SQLite for localized data persistence, and Streamlit for real-time dashboards, the framework eliminates proprietary vendor lock-in. Furthermore, its modular design is explicitly optimized for affordable edge hardware like the Raspberry Pi. This demonstrates how openecosystem tools can be integrated to build resilient, enterprise-grade localized solutions without necessitating expensive, continuous cloud connectivity. This edge-deployed system directly addresses critical disparities in global healthcare access. By fusing multimodal physiological data locally, the framework provides automated, real-time patient risk stratification in resourceconstrained or remote environments where internet bandwidth is often unreliable. The system’s offline fallback capabilities ensure continuous clinical operations regardless of local infrastructure failures. Ultimately, by delivering high-accuracy, rapid diagnostic insights to the physical edge at a fraction of the cost of traditional medical systems, this project empowers under-resourced clinics to efficiently prioritize critical patients, reducing mortality risks and advancing health equity in underserved communities.
| Agree to Privacy Policy and Notice | I agree |
|---|---|
| In Person Attendance | Remote |
| Please confirm that there are included headshots of all speakers in their profiles | Yes |