README.md

metadata

title: Ocr
emoji: 👀
colorFrom: blue
colorTo: pink
sdk: streamlit
sdk_version: 1.38.0
app_file: app.py
pinned: false

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

Libraries Required

streamlit torch transformers Pillow easyocr huggingface_hub requests re

Setup & Installation

• Python 3.8+
• pip (Python package installer)

Features

• Dual Language OCR: Supports text extraction from images containing both Hindi and English text.
• Fallback System: Automatically switches to EasyOCR if the GOT model fails.
• Image Upload: Users can upload image files (JPEG, PNG) for OCR processing.
• Text Search: Allows users to search for keywords within the extracted text and highlights the matching results.
• Simple UI: Easy-to-use interface for non-technical users.
• Evaluation Criteria

Accuracy

• The application extracts text from both Hindi and English sections of the image with high accuracy, using either GOT-OCR2_0 or EasyOCR.

Functionality

• Handles image uploads (PNG, JPEG).
• Extracts text from images and displays it in plain text format.
• Allows users to search for keywords within the extracted text, with the matching results highlighted.
• User Interface
• A simple, intuitive interface built with Streamlit.
• Provides clear instructions and feedback (e.g., displaying progress during OCR, error handling).
• Responsive layout to fit different screen sizes.

Deployment

• The application can be deployed online via a cloud service like Streamlit Cloud, Heroku, or any other hosting platform that supports Python applications.

Clarity

• Well-structured and easy-to-follow code.
• Clear comments and logical flow for each function.
• User-friendly interface with concise instructions for end-users.

Completeness

• All functionalities are implemented and operational.
• The application works as expected for Hindi and English OCR and includes keyword search.
• Prerequisites

cd got-ocr-app

streamlit run app.py