Curvelets obey parabolic scaling (width ≈ length²), making them significantly more efficient than wavelets for images with curved edges. For side-scan sonar — hull shadows, seafloor targets, debris — this means better denoising and edge enhancement at the same coefficient budget.
The personal context: I'm a sailor. I worked in Congress as a special assistant to a congressman and in campaign consulting — I've seen firsthand how bureaucratic systems slow down or block tools that could save lives. In 2024, Rosa and Charlie Brown went missing at sea. The Coast Guard declined to run a backdrift analysis. I know exactly what tools could have helped, and I know exactly why they weren't available to the volunteers who needed them.
I was diagnosed with ADHD, dyslexia, and dysgraphia as an adult. I build software to compensate for gaps I've experienced myself — and this crate is the signal-processing foundation for CESARops: a full open-source SAR coordination platform with drift prediction, search planning, and sonar analysis, free for civilian and volunteer SAR teams permanently.
(Separate project: Wayfinder — an ADHD-friendly file naming and organization tool I originally built for myself, being expanded for school-age kids.)
Looking for: Rust contributors, DSP specialists, SAR professionals, technical writers. Issues: https://github.com/festeraeb/nauticuvs/issues Sponsor: https://github.com/sponsors/festeraeb