Benchmark RadarFloodReasonBench
FloodReasonBench is a benchmark for vision-language model reasoning segmentation in flood response scenarios. It introduces FloodResponseSeg, a flood-specific dataset, and evaluates pipelines under lightweight visual encoding, split inference, and compressed representations. It also measures accuracy, latency, energy, and communication tradeoffs on an embedded platform.
- Released
- 2026-08-15
- Readiness
- Paper only
- Primary field
- Robotics & Autonomous Systems
Why it matters
Reasoning segmentation for flood response has domain-specific constraints, and this benchmark characterizes model performance and system-level tradeoffs at the edge, which could inform deployment decisions for resource-constrained platforms.
Motivation
Reasoning segmentation enables vision-language models (VLMs) to translate mission-relevant language requests into pixel-level visual grounding, offering a natural perception interface for embodied agents.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.