.png&w=256&q=75)
🤝 Top Collaborators
.png&w=640&q=75)
.png&w=640&q=75)
🤓 Latest Submissions
DisasterIQ
DisasterIQ — see damage, prioritize relief, save lives. After a tornado, flood, or earthquake, the bottleneck in emergency response is not imagery — satellites deliver that within hours. The bottleneck is triage: nobody can tell coordinators which streets to reach first. Analysts hand-compare before/after tiles for days while search-and-rescue teams wait. DisasterIQ turns a before/after satellite pair into a ranked deployment plan in about twelve seconds. Upload two images (or load an xBD demo pair) and the pipeline co-registers the frames with FFT phase correlation, cancels lighting differences, and isolates genuine structural change from noise. Connected-component analysis extracts damaged building footprints, classifies them by severity — minor, major, destroyed — and scores a zone grid to produce a ranked priority list. The dashboard renders a damage overlay, a zone map with real-world coordinates, and a plain-language situation brief written for a field coordinator, exportable as a PDF field report. Every request runs on AMD compute, and you can check it rather than take our word for it. The damage pipeline — alignment, change detection, building extraction, zone scoring — executes on an AMD EPYC 7R13: curl https://disasteriq-backend.onrender.com/compute. The AI situation brief is generated by gpt-oss-120b on Fireworks AI, which serves inference on AMD Instinct GPUs.
13 Jul 2026
