RuihengZhang/IFSOD-dataset resource

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1005 stars · 3 forks observed · 2026-09-03

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

IFSOD is a benchmark dataset for infrared few-shot object detection, containing 4,815+ images with 23,000+ annotations across 18 categories and 12 scenarios. It ships with predefined base/novel class splits following the TFA few-shot detection protocol.

Use cases

• train few-shot object detection models on infrared imagery
• benchmark infrared object detection algorithms
• evaluate novel-class generalization with 10-shot splits
• research military and civilian target detection in thermal images
• compare against M3FD and RGB-T234 infrared datasets

When to choose

• you need labeled infrared images for few-shot detection research
• you want standardized base/novel splits matching the FSDet setting
• your work targets thermal/night-vision object detection benchmarks

When to avoid

• you need large-scale RGB or RGB-T paired data
• you require a permissive open-source license for commercial use
• you need video or multimodal infrared-visible datasets

Facets

dataset · maturity active

computer-vision machine-learning data-science computer-vision machine-learning deep-learning cross-platform infrared-imaging few-shot-learning object-detection benchmark-dataset thermal-imaging

Member repositories

For agents

markdown · JSON · MCP: product_card(name="RuihengZhang/IFSOD-dataset")