Awesome Computational Imaging

Created on April 30, 2025

2025   ·   computational-imaging   diffusion-models   inverse-problems   nerf   ·   blog

I created and maintain the open resource:
🌐 Awesome Computational Imaging

A curated learning and implementation series on modern computational imaging — bridging physical models and deep learning.

🔍 Why This Project?

As an undergraduate in Optoelectronics with a passion for deep learning, I found it difficult to find unified resources for topics like:

  • Inverse problems in scientific imaging
  • Physics-informed neural networks
  • Neural field representations (e.g., NeRF, SIREN, FFN)
  • Diffusion-based image restoration
  • Plug-and-play methods in computational optics

Most tutorials were either:

  • Too theoretical, or
  • Lacked code-level explanations

📚 What’s Included

The website contains modular learning units, each with:

  • 📖 Theory: Core concepts & physical models
  • 💡 Code: Simple yet extensible PyTorch implementations
  • 🧪 Applications: Imaging tasks like deblurring, phase retrieval, envelope removal, and more
  • 🧵 Trends: From NeRF and SIREN to diffusion-based inverse solvers (e.g., DPS, ILVR)

🧩 Topics Covered (so far)

  • ✅ Implicit Neural Representations
    • SIREN
    • NeRF
    • Fourier Feature Networks (FFN)
  • ✅ Diffusion Models for Inverse Problems
    • DPS
    • RePaint
    • ILVR
    • ControlNet-based occlusion recovery
  • ✅ Physics-based Forward Models
    • Atmospheric scattering
    • Holography & phase retrieval
  • ✅ Plug-and-Play (PnP) Priors
    • Denoising priors (DnCNN, SwinIR)
    • Real-ESRGAN + physics consistency

🌟 Goals

The vision of this project is to:

Make computational imaging accessible to researchers, students, and engineers
by providing end-to-end understanding — from physics to code to results.

I hope this project can also serve as a portfolio and learning hub for students entering the field.

💬 Get Involved

If you’re working on related topics or want to contribute:

📌 Last updated: August 2025
✍️ Maintained by Xueqing Gao