PyDoppler

PyDoppler is a Python prototype inspired by HoloDoppler, designed to explore the benefits of GPU acceleration (CUDA) for data processing.

Goal

Test a faster and more flexible alternative to the existing post-processing, without rebuilding a full application.

What it does

The pipeline reconstructs a vascular flow image from a raw interferogram:

Input — interferogram	Output — vascular reconstruction

How it works

Simple processing pipeline:

• Spatial transformation
• Temporal transformation
• Data accumulation
• Moving average

Can run on:

• CPU (multithreaded)
• GPU (CUDA)

Results

• Processing time reduced: ~5 min → ~30 s per file
• Results obtained without heavy optimization

Why this prototype?

• Evaluate performance gains with CUDA
• Test a more modular Python approach
• Serve as a base for a possible future version

Tech

• Python
• CUDA
• Signal processing

---

⚠️ Experimental prototype:
This project is not a complete application, but a working base to test ideas and performance.