👁️ Universal Access AI

A Next-Generation Desktop Assistant for the Visually Impaired

Universal Access AI is an intelligent accessibility tool designed to help users with vision impairments navigate their digital and physical environments independently. By combining Computer Vision, Optical Character Recognition (OCR), and Generative AI, it acts as a digital pair of eyes that can see, read, and converse in real time.

---

🚀 Key Features

• 🗣️ Full Voice Control
  Completely hands-free operation. Users can issue voice commands and receive audio feedback.

• 🔍 Object Detection
  Locates everyday items (e.g., “Find my phone”, “Where are my keys?”) and provides spatial guidance.

• 📷 Smart OCR Scanner
  Reads text from physical documents, books, and screens instantly.

• 👁️ Scene Description
  Analyzes the live camera feed and describes the environment in natural language
  (e.g., “I see a person sitting at a desk with a laptop”).

• 🤖 AI Companion
  Integrated with Llama 3 (via Groq) to answer general questions and provide intelligent assistance.

• ⚡ Zero-Latency Mode
  Uses multi-threading to ensure camera, voice, and AI modules run smoothly without UI freezing.

---

🛠️ Installation Guide

Prerequisites

• Python 3.10 or newer
• Tesseract OCR (required for text reading features)
  https://github.com/UB-Mannheim/tesseract/wiki

---

1️⃣ Clone the Repository

git clone https://github.com/YOUR_USERNAME/Universal-Access-AI.git
cd Universal-Access-AI

2️⃣ Install Dependencies
pip install -r requirements.txt

3️⃣ Download Required AI Models (Critical Step)

This project uses YOLOv3 for object detection. Due to file size limits, the weights must be downloaded manually.

Download yolov3.weights (~237 MB)

Place the file in the main project directory (same folder as main.py)

Ensure the following files are present:

yolov3.weights

yolov3.cfg

coco.names

4️⃣ Install Tesseract OCR

For the Read Text feature to work:

Download and install from:
https://github.com/UB-Mannheim/tesseract/wiki

Default install path:
C:\Program Files\Tesseract-OCR

The application automatically detects this path.

5️⃣ Run the Application
python main.py

🌟 Key Benefits

Accessibility First
Designed specifically for visually impaired users to navigate the world independently.

Cost-Effective
Runs on a standard laptop with a webcam—no expensive hardware required.

Privacy-Focused
Video processing and object detection run locally on the device.

Multimodal Interaction
Combines voice, vision, and text into a single seamless interface.

👨‍💻 Project Built By

Bhavan Kothalanka
B.Tech 3rd Year Student | AI & Data Science Enthusiast

I am a passionate developer focused on leveraging Artificial Intelligence to solve real-world problems. With a strong foundation in Data Science and Computer Vision, I built Universal Access AI to bridge the gap between technology and accessibility. I enjoy exploring how Large Language Models and Edge AI can be combined to create impactful, human-centered software.