Skip to content
Book Detection for Visually Impaired
Feb 2024 - Jun 2024

Book Detection for Visually Impaired

Real-time OCR-to-speech system for accessibility

01. The Challenge

Visually impaired individuals lack accessible tools for reading physical books.

02. The Solution

Developed real-time computer vision system with OCR and text-to-speech for book accessibility.


Overview

Access to physical books remains a challenge for the visually impaired. This computer vision system runs on low-power devices to detect book pages, unwarp them, and read the text aloud.


Pipeline

  • **Detection:** YOLOv8 detects the book object.
  • **Segmentation:** Extracts the page area.
  • **Dewarping:** Corrects the perspective and page curl.
  • **OCR:** Tesseract extracts text.
  • **TTS:** Converts text to speech.

    • System Architecture

    OpenCV for book detection
    PyTesseract for OCR
    pyttsx3 for text-to-speech
    Real-time processing pipeline

    Project Links

    Technologies

    OpenCVPyTesseractpyttsx3Python

    Key Impact

    Real-time book reading
    Accessibility enhancement
    Low-latency processing