👁️ SymLens
Through the lens of your symptoms."
Built for the HP AI Studio Hackathon 2025 – Healthcare Category
🔍 Overview
SymLens is an on-device AI-powered visual symptom simulator that allows patients to turn hard-to-describe visual disturbances—like migraine aura, floaters, photopsia, and visual snow—into realistic images. Built using HP AI Studio’s compact vision-language models, SymLens enables clearer, faster communication between patients and clinicians, all while preserving privacy with fully local inference.
🩺 Problem Statement
Many patients, especially in neurology and ophthalmology, struggle to explain what they see. Whether it’s flickering lights, translucent blobs, or static-like vision, these experiences are often:
- Misunderstood by clinicians
- Poorly documented in medical records
- Emotionally distressing for patients
No current tool allows personalized, visual communication of subjective symptoms in a fast, secure, and interactive way.
💡 Solution
SymLens bridges the gap between symptom and sight. Users simply describe what they see, and SymLens generates realistic, AI-powered visual renderings. These can be exported, refined, or used in clinical visits—helping doctors diagnose more accurately and giving patients the confidence that they’re being understood.
🎯 Key Features
- 📝 Free-Text Symptom Input — Write or speak what you see
- 🧠 On-Device AI Inference — Powered by HP AI Studio’s optimized LLMs + vision models
- 🎨 Text-to-Image Visualizations — Generated locally using fine-tuned diffusion models
- 🔁 Interactive Adjustments — Compare multiple renders, tweak blur, floaters, opacity
- 📤 Exportable Visuals — Save/share images with your provider or for tracking over time
⚙️ How It Works
- Describe your symptom (e.g., “like raindrops on a foggy window”).
- HP AI Studio processes your prompt using TinyVLM locally.
- Image is generated via an on-device, quantized diffusion model.
- You review and refine visual accuracy.
- Export or save your visual for a clinical visit.
🧰 Technology Stack
- AI Platform: HP AI Studio
- Vision-Language Models: TinyVLM / BLIP (edge-optimized)
- Image Generator: Quantized Stable Diffusion
- Frontend: React.js + HTML5 Canvas
- Backend: FastAPI (Python)
- App Container: Electron (for cross-platform desktop app)
- Storage: Local SQLite DB (privacy-focused)
📂 Public Models & Tools Used
- TinyVLM / BLIP – Text-image model
- Stable Diffusion – Modified for symptom-specific visuals
- HP AI Studio SDK – Edge deployment and tuning
🚀 Future Roadmap
- 📱 Mobile support with offline inference
- 🧠 Emotional overlay analysis for trauma-related hallucinations
- 🧾 Integration with EHRs (via secure export)
- 🌍 Multilingual & speech-to-image capabilities
🤝 Team
Aishwarya Jakka
Full-Stack Developer | MBA (Applied AI) | Healthcare Innovator
- GitHub: aishwaryajakka
- LinkedIn: linkedin.com/in/aishwaryajakka
SymLens helps patients be seen, not just heard—through the lens of what they truly experience.

Log in or sign up for Devpost to join the conversation.