HumanSense AI: Turning Human Signals into Real-Time Insights 🚀
This is a submission for the Presage SmartSpectra SDK Challenge.
💡 What I Built
Meet HumanSense AI — a camera-based human signal monitoring desktop application built with Electron, Node.js, and the Presage SmartSpectra SDK.
The idea behind HumanSense AI is simple: make complex human physiological signals easier to visualize and understand through a modern, interactive desktop experience.
The project focuses on four key signal categories:
- ❤️ Pulse: Heart rate measurements.
- 🌬️ Breathing Rate: Respiratory rate insights.
- 📊 HRV: Heart rate variability.
- 😊 Facial Expressions: SDK-supported facial expression analysis.
I wanted to explore how camera-based sensing technology could create new possibilities for human-computer interaction.
🎥 Demo
Watch HumanSense AI in action:
[Add your YouTube or public demo video link here]
The demo should showcase:
- Application startup.
- Camera permission and user consent.
- SmartSpectra SDK initialization.
- Live signal measurements.
- Real-time dashboard updates.
- Starting and stopping a measurement session.
💻 Source Code
GitHub Repository: [https://github.com/Bilal0080/HumanSense-AI---SmartSpectra-Vitals]
The project uses:
- Electron for the desktop application.
- Node.js for application-side functionality.
- Vite for development and building.
- Presage SmartSpectra SDK for human signal processing.
- HTML, CSS, and JavaScript for the interface.
The goal is to maintain a clean project architecture with secure communication between the Electron main process, preload layer, and renderer.
🧠 How It Responds to Human Signals
HumanSense AI is designed to use webcam input with the Presage SmartSpectra SDK to obtain supported physiological measurements.
The intended workflow is:
- The user opens HumanSense AI.
- The application requests permission to access the camera.
- The user provides consent before measurement begins.
- The application initializes the SmartSpectra SDK.
- Supported signals are processed and received.
- Available measurements are displayed through dashboard cards.
- The user can stop the session whenever they want.
The dashboard is designed to make signal information easy to understand without overwhelming the user with technical details.
An important part of this project is distinguishing actual SDK measurements from unavailable signals. HumanSense AI should never present fabricated readings as real measurements.
🔐 Consent and Data Handling
Privacy and user consent are important parts of HumanSense AI.
The application is designed around the following principles:
- Camera access requires user permission.
- Users should understand why camera access is needed.
- Measurement should only begin after user consent.
- Users can stop a measurement session.
- The application is not intended for covert monitoring.
- Facial expression estimates must not be interpreted as definitive evidence of a person's actual emotions.
- HumanSense AI does not claim to diagnose, treat, or prevent medical conditions.
Data storage, transmission, retention, deletion, and SDK telemetry depend on the verified SDK configuration and application implementation. These behaviors must be documented accurately before public release.
HumanSense AI is an experimental software project, not a medical device.
🛠️ Challenges and Learning Experience
One of the most interesting challenges was integrating a native Node.js SDK into an Electron application.
The project also involved understanding:
- Electron main-process and renderer architecture.
- Native dependencies and platform compatibility.
- Secure IPC communication.
- Camera permissions.
- SDK initialization and measurement handling.
- Vite configuration for Electron projects.
- Privacy-conscious application design.
This project has been a valuable learning experience in connecting native SDK technology with a modern desktop interface.
🚀 What's Next?
I would like to continue improving HumanSense AI by exploring:
- More detailed real-time signal visualizations.
- Better measurement quality indicators.
- Improved dashboard interactions.
- Session summaries, subject to privacy safeguards.
- Better error handling and SDK diagnostics.
- Additional SmartSpectra-supported capabilities.
💬 Let's Connect!
I would love to hear feedback from the developer community.
Feel free to leave your questions and suggestions in the comments!
