A multi-task face model with four underage thresholds and a frozen FaRL backbone, trained with age-balanced resampling and an age gap, improves underage detection on new benchmarks ASORES-39k and ASWIFT-20k.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Underage Detection through a Multi-Task and MultiAge Approach for Screening Minors in Unconstrained Imagery
A multi-task face model with four underage thresholds and a frozen FaRL backbone, trained with age-balanced resampling and an age gap, improves underage detection on new benchmarks ASORES-39k and ASWIFT-20k.