Back to work
DeepFract
Orchestrated AI models for fractal image compression — high ratio and detail retention in a short encode time. FastAPI backend, Flutter mobile client.
- Python
- FastAPI
- Flutter
- PyTorch
- Computer Vision

Overview
DeepFract is an AI-enhanced fractal image compression system. Instead of relying on a single model, it orchestrates several specialized networks that work together — residual CNNs, attention (CBAM, attention-gated U-Net), and quad-tree partitioning — so encoding stays fast while compression ratio and visual detail stay high.
Hard parts
- 01Coordinating multiple specialized models so they improve ratio and quality without blowing up encode time.
- 02Designing attention that surfaces structural self-similarities useful for fractal block matching.
- 03Bridging a PyTorch inference backend with a responsive Flutter mobile client.