Video & Image OCR Tool

A Streamlit-based web application to extract text from uploaded images and videos using EasyOCR.

Features

• Image OCR: Upload an image (PNG, JPG, JPEG) and extract text.
• Video OCR: Upload a video (MP4, MOV, AVI) and extract text from frames at a specified sampling rate.
• Video ASR: Extract text from the audio channel of videos using automatic speech recognition.
  • Supports local inference using Hugging Face models (e.g., OpenAI Whisper, Wav2Vec2).
  • Uses GPU acceleration if available.
• Multi-language Support: Support for English, Chinese (Simplified), French, German, Spanish, Japanese, Korean, and Dutch.
• Downloadable Results: Download the extracted text as a .txt file.

Installation

1. Clone the repository (if applicable) or ensure you have the project files.
2. Install the required dependencies:

pip install -r requirements.txt

Note: For video audio extraction, ffmpeg is required. moviepy usually handles this, but if you encounter issues, ensure ffmpeg is installed on your system.

Running the App

Run the Streamlit application using the following command:

streamlit run app.py

Usage

1. Select Languages: Use the sidebar to select one or more languages for OCR. The default is English.
2. Upload File: Upload an image or video file.
3. Extract Text:
   • For Images: The extracted text will be displayed immediately after clicking "Extract Text".
   • For Videos: You can adjust the sampling rate (seconds per frame) in the sidebar. Click "Extract Text from Video" to process.
4. Download: Click the "Download Text" button to save the results.

Notes

• First run might be slow as EasyOCR downloads the necessary models.
• Video processing can be time-consuming depending on the video length and sampling rate.