CoolFace
Apppublic

K2-k2/AIVision

sourceHugging Faceupdated 6mo agoView on Hugging Face
0likes
App README

🤖 The Blind Assistant

The Blind Assistant is a specialized multimodal AI application designed to provide accessibility for the visually impaired. Using Google Gemini 2.5 Flash for vision-to-text and gTTS (Google Text-to-Speech), it converts visual surroundings into clear, safety-focused audio descriptions.

🌟 Overview

This tool serves as an artificial eye. When a user uploads an image, the AI analyzes the scene—identifying objects, hazards, and spatial context—and generates both a written transcript and an audible narration. It is built to prioritize clarity and conciseness, specifically for individuals who rely on auditory feedback.

🛠️ Key Features

  • —Multimodal Vision: Leverages Gemini 2.5 Flash to "see" and interpret complex visual scenes.
  • —Safety-First Prompting: Optimized to highlight potential hazards and key environmental objects.
  • —Text-to-Speech (TTS): Converts descriptions into natural-sounding audio instantly.
  • —Audio Downloads: Users can download the MP3 narration for offline reference.
  • —Clean Interface: A minimalist, accessible UI built with Streamlit.

🚀 Deployment on Hugging Face Spaces

To host this on Hugging Face:

  1. 1.SDK: Select Streamlit during the Space creation.
  2. 2.Secret Keys: Add your API key in Settings > Variables and Secrets as GOOGLE_API_KEY_N.
  3. 3.Requirements: Ensure your requirements.txt includes:
text
    streamlit
    Pillow
    gTTS
    google-generativeai

🏗️ The Tech Stack

  • —Frontend: Streamlit
  • —Vision Model: Google Gemini 2.5 Flash
  • —Image Processing: PIL (Pillow)
  • —Voice Synthesis: gTTS (Google Text-to-Speech)

💻 Local Setup

bash
# 1. Clone the repository
git clone [https://huggingface.co/spaces/K2-k2/AIVision
](https://huggingface.co/spaces/K2-k2/AIVision
)

# 2. Install dependencies
pip install -r requirements.txt

# 3. Configure your API key
# Windows: set GOOGLE_API_KEY_N=your_key
# Mac/Linux: export GOOGLE_API_KEY_N="your_key"

# 4. Launch the application
streamlit run app.py

Developed by Kesar Agarwal an aspiring AI Engineer dedicated to creating impactful, accessible AI solutions for real-world challenges.

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference