README.md

metadata

title: Manga Proj
emoji: 📈
colorFrom: green
colorTo: pink
sdk: streamlit
sdk_version: 1.42.2
app_file: app.py
pinned: false

Manga Proj

This Streamlit app processes manga images by translating text within speech bubbles using a YOLO model for text detection and Tesseract OCR for text recognition.

Setup

To ensure the app runs correctly, follow these setup instructions:

1. Install Dependencies:

   • The app requires several Python packages listed in requirements.txt.
   • Additionally, Tesseract OCR needs to be installed on the system.

2. Run Setup Script:

   • Execute the setup.sh script to install Tesseract OCR:

     bash setup.sh
     

Usage

1. Upload Files:

   • Upload a manga image, a CSV file containing translations, and a font file.

2. Run the Pipeline:

   • Click the "Run Full Pipeline" button to process the image.
   • The app will display the processed image with translated text.

Features

• Text Detection: Uses a YOLO model to detect text regions in the image.
• Text Recognition: Utilizes Tesseract OCR to recognize and extract text.
• Translation: Matches extracted text with translations provided in a CSV file.
• Image Processing: Cleans text regions and renders translated text back onto the image.

Dependencies

• Streamlit
• Ultralytics (YOLO)
• Keras OCR
• Sentence Transformers
• Scikit-learn
• Pillow
• OpenCV
• Pytesseract

Acknowledgments

• Thanks to the open-source community for providing the tools and libraries used in this project.

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