Easylens

EasyLens is an open-source undergraduate thesis project built to empower visually impaired individuals with safer, smarter, and more independent mobility through AI.

Unlike traditional navigation apps, EasyLens combines real-time object detection, voice-guided navigation, wearable smart glasses, and an intelligent AI assistant into one accessibility platform. Our goal is to make everyday travel more confident by helping users recognize obstacles, understand their surroundings, and navigate using natural voice interactions.

At the heart of EasyLens is our custom fine-tuned MobileNetV2 object detection model, trained, combining 400+ classifying object include google ml kit, combining with gemma2 Local LLM  and optimized specifically for assistive navigation. Running directly on-device, it delivers fast, low-latency obstacle detection without relying on constant cloud processing, making it suitable for real-world use even in challenging environments.

Key highlights include:

- Custom Fine-Tuned MobileNetV2 AI Model for real-time obstacle detection

- Integration with ESP32-CAM smart glasses for wearable computer vision

- AI-powered voice navigation with turn-by-turn guidance

- Filipino & English voice assistant with multiple voice personas

- OCR-powered text reading for signs, labels, and printed materials

- SOS emergency system with instant SMS alerts

- Designed with accessibility and inclusive technology at its core

- 100% Open Source to encourage AI researcher, collaboration, and innovation in assistive AI

EasyLens isn't just another AI app it's a AI research-driven accessibility platform that demonstrates how edge AI, mobile computing, and computer vision can improve the daily lives of people with visual impairments.

(Note:Currently under active development and validation as our thesis project. Public deployment is coming soon.)

We hope EasyLens inspires more developers, researchers, and accessibility advocates to build technology that creates meaningful social impact.

Developer: Arron Kian Parejas
Hardware: Jian Kalel Marquez
Documentation and UIX: Graciella Mhervie Jimenez, Jenica Sarah Tongol