docs/README.en.md

metadata

license: mit
title: Pic-to-Header
sdk: streamlit
emoji: 🐨
colorFrom: blue
colorTo: purple
pinned: false
app_file: pic_to_header/app.py

Pic-to-Header

Pic-to-Header is a Python application that generates header images using a mask image and an input image. Version 0.2.0 has been released.

🚀 Project Overview

Pic-to-Header is a Streamlit application that allows you to easily generate and download header images using a mask image and an input image. It's also available on PyPI. It also supports a command-line interface (CLI).

🎥 Demo Video

https://github.com/user-attachments/assets/9306ad92-9a2e-45d7-9004-0eba9c0e1207

🤗 Demo Site

✨ Key Features

• Upload mask and input images
• Generate header images
• Preview and download generated header images
• Image processing using a command-line interface (CLI)
• Mask management features
• Extended image processing logic
• Image processing functionality via Python script
• Addition of README page and release notes page to the Streamlit app

🔧 Usage

Method 1: Installation from PyPI

Pic-to-Header is available on PyPI. You can install it using the following command:

pip install pic-to-header

Method 2: Installation from Source

1. Clone the repository:

   git clone https://github.com/Sunwood-ai-labs/pic-to-header.git
   cd pic-to-header
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

Streamlit Web Application

1. Start the Streamlit application:

   streamlit run pic_to_header/app.py
   

2. Access the URL displayed in your browser.
3. Upload the input image and mask image.
4. Click the "Generate Header Image" button.
5. Preview the generated header image and download it if necessary.

Command-Line Interface (CLI)

You can also process images using the CLI:

pic-to-header input_image.png mask_image.png output_image.png

Example:

pic-to-header assets/sample.png assets/mask.png output_image.png

Using a Python Script

The pic_to_header package can also be used directly within Python scripts. Here is an example:

from pic_to_header.core import process_header_image

# Specify the paths to the input image, mask image, and output image
input_image_path = "path/to/input_image.png"
mask_image_path = "path/to/mask_image.png"
output_image_path = "path/to/output_image.png"

# Generate the header image
result = process_header_image(input_image_path, mask_image_path, output_image_path)

print(f"Header image generated: {result}")

This script works as follows:

1. Imports the process_header_image function from the pic_to_header.core module.
2. Specifies the paths to the input image, mask image, and output image.
3. Calls the process_header_image function to generate the header image.
4. Displays the path to the generated image.

Running this script will generate the header image at the specified output path.

💻 Development

• pic_to_header/core.py: Contains the main image processing functionality.
• pic_to_header/app.py: Provides the web interface using Streamlit.
• pic_to_header/cli.py: Provides the command-line interface.

📦 Installation Instructions

See the "Usage" section above.

🆕 What's New (v0.2.0)

• Mask Management Feature Implementation: Mask images can now be obtained from presets, URLs, and uploaded files. A preset mask management feature has also been added.
• Extended Image Processing Logic: Automatic adjustment of the number of channels for images and mask images, mask image resizing, mask transparency adjustment, and the ability to return processing results as PIL Images have been added.
• Added Python Script Image Processing Functionality and Updated Documentation: The usage of the process_header_image function is explained step-by-step.
• UI Improvements: Streamlit wide mode setting, addition of mask image acquisition methods (presets, URL, upload), mask transparency adjustment slider, support for multiple image uploads, changes to the display method of processing results, addition of a download button, and enhanced error handling have been implemented.
• Added links to Hugging Face Spaces demo site and demo video.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.