Stroke Detection App This web-app prototype is designed to detect early signs of stroke using device sensors, with a focus on accessibility and real-time offline analysis. The goal is to provide users—especially those in high-risk groups or remote areas—with a quick, intelligent screening tool that compares current health data with their own baseline to identify signs of stroke onset.

How It Works Upon registration, users record their baseline data through: Camera: Captures facial structure while smiling to assess natural symmetry. Microphone: Records voice samples to analyze speech clarity. Accelerometer: Collects data from arm movements to monitor for potential weakness or drift. This information is securely stored locally (no internet needed), forming a unique health profile for each user. During a stroke test, the app requests the same inputs and performs real-time comparisons to flag: Facial Drooping (first sign): Detected via facial landmark symmetry checks. Slurred Speech: Identified by changes in voice articulation and cadence. Muscle Weakness: Detected by analyzing arm stability using sensor data. If deviations from baseline exceed defined thresholds, the app alerts the user and recommends immediate medical attention.

Key Features Offline-first design: Works without internet after initial setup. Privacy-first: All data is stored and processed locally on the user’s device. Simple UI: Designed for elderly and non-technical users. Extensible: Modular architecture allows easy integration of cloud services or ML models.

Status Currently in prototype stage. This was made for Mission: Brain Hackathon 2025.