Benchmark RadarAI BENCHMARK PROFILE
FunPiQ
FunPiQ is a benchmark for pixel-level fundus image quality assessment with pixel-level annotations for anatomical visibility, using a three-class scheme (good, usable, bad). It also introduces EFIQA-CP, an explainable method trained with pseudo-labels.
- Released
- 2026-06-24
- Readiness
- Runnable
- Primary field
- Health & Life Sciences
Why it matters
Existing FIQA benchmarks provide only image-level labels, limiting localized degradation analysis. FunPiQ enables task-agnostic explainable quality evaluation through pixel-level ground truth and offers a public dataset for method comparison.
Motivation
Color fundus photography (CFP) is the most common ophthalmic imaging modality for large-scale screening.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.