PropFix: AI-Assisted Photo Editing Electron Application

PropFix is a cross-platform desktop application that pairs a modern Electron/Node.js user interface with a Python back end to deliver intelligent photo-editing tools. The goal is to demonstrate how computer systems engineering, data science, and modern software architecture can be combined in a portable, product-grade application. As a passion project, the repository emphasizes reproducibility, documentation, and careful version control so it can serve as part of a professional portfolio.

---

Table of Contents

1. Motivation
2. Features
3. Technology Stack
4. Architecture
5. Installation
6. Usage
7. Data-Science Details
8. Model Training & Integration
9. Version Control & Contribution Guidelines
10. License
11. Acknowledgements

---

Motivation

Modern photo editors often rely on a mix of manual sliders and presets. As a computer systems engineer with a data-science focus, I wanted to explore how far AI can automate high-quality edits while still keeping the user in control. Electron makes it straightforward to build cross-platform desktop apps, and Python brings mature image-processing and ML tooling. PropFix integrates them so the UI stays responsive while heavy lifting happens in a separate Python process.

---

Features

PropFix is designed to evolve into a full-fledged AI-assisted photo editor. The current roadmap includes:

• Photo management — load single images or folders, display thumbnails, and organize your work.
• Automatic enhancements — color balance, white balance, noise reduction via pre-trained deep-learning models.
• Object detection and segmentation — identify regions (e.g., person/parts) for selective, region-aware edits.
• Generative editing — inpainting and style transfer via modern diffusion/GAN-based approaches.
• Undo/Redo and history — keep a timeline of edits so you can revert or replay.
• Data export — save PNG/JPEG/TIFF and export JSON metadata describing the edit pipeline.

UI niceties already implemented include a log pane, “Reset All,” sensible defaults, and responsive zoom controls (Fit/Fill/1:1). The Python side produces preview images (see server/outputs/).

---

Technology Stack

PropFix intentionally uses a hybrid stack:

Layer	Technology / Libraries	Rationale
User Interface	Electron (HTML/CSS/JavaScript)	Portable, modern desktop UI; access to filesystem and native menus while rendering a web-quality interface.
Application Logic	Node.js	Coordinates renderer events, filesystem I/O, and calls to the Python back end over HTTP.
Back End	Python 3 (Flask, Pillow/OpenCV, NumPy, etc.)	Robust image processing, segmentation/warping, and room for ML models. Exposed as an HTTP server so the UI remains responsive.
Data Science	Deep-learning models (U-Net/Mask-RCNN, ESRGAN/DnCNN, diffusion, etc.)	Segmentation, super-resolution/denoising, inpainting, style transfer—extensible via Python modules.
Version Control	Git + Conventional Commits	Clean history and semantic-style commit messages.

---

Architecture

The application uses a simple, testable separation of concerns:

1. Renderer (Electron)
   Renders the UI, gathers user inputs (sliders, toggles), and issues HTTP requests to the local Python server.
   Example request:

   // classify → returns segments; warp → applies geometry and photo controls
   const res = await fetch('http://127.0.0.1:5001/process', {
     method: 'POST',
     headers: { 'Content-Type': 'application/json' },
     body: JSON.stringify({
       mode: 'warp',                     // or 'classify'
       input: 'C:\\path\\to\\image.png', // absolute path from the UI
       controls: {
         geometry: { scaleX: 1.02, scaleY: 0.98, rotate: -1.5, offsetX: 0, offsetY: 0 },
         photo:    { contrast: 1.0, exposure: 0.0, saturation: 1.0, strength: 0.75 }
       }
     })
   });
   const data = await res.json();