Intelligent OCR and Text Analysis Tool

Description

An advanced application that performs Optical Character Recognition (OCR) on images and PDFs, extracts text, and provides a question-answering interface based on the extracted content. It leverages machine learning models and modern NLP techniques to enable users to interactively query their documents.

Techniques and Tools Used

• Streamlit: For building the interactive web application.
• PyPDF2: To read and extract text from PDF files.
• Pillow (PIL): For image processing and manipulation.
• OCR Module: Custom module (ocr_module.py) for performing OCR on images.
• RAG Module: Custom module (rag_module.py) implementing Retrieval-Augmented Generation for processing queries.
• Transformers: HuggingFace library for loading pre-trained models.
• SentenceTransformers: For generating sentence embeddings.
• PyTorch: Deep learning framework underpinning the ML models.

Features

• Upload Images or PDFs: Accepts multiple image formats and PDFs for text extraction.
• Perform OCR: Extracts text from images using the perform_ocr function.
• Text Analysis: Enables users to ask questions about the extracted text using the process_query function.
• Custom Styling: Utilizes custom CSS and JavaScript for an enhanced UI/UX.

Code Snippets

Loading Custom CSS

def load_css():
    with open('static/styles.css') as f:
        st.markdown(f'<style>{f.read()}</style>', unsafe_allow_html=True)

Handling File Uploads

def handle_file_upload(uploaded_file):
    if uploaded_file.type == "application/pdf":
        pdf_reader = PyPDF2.PdfReader(uploaded_file)
        text = ""
        for page in pdf_reader.pages:
            text += page.extract_text()
    else:
        image = Image.open(uploaded_file)
        text = perform_ocr(image)
    return text

Processing User Queries

def get_answer(query, context):
    answer = process_query(query, context)
    return answer

Contact

For inquiries or feedback:

• Email: rayyanahmed265@yahoo.com
• LinkedIn: Rayyan Ahmed
• GitHub: Rayyan9477