MobileNetV2 + CBAM attention + two-phase layer unfreezing reaches 92.4% accuracy on a curated 68-identity LFW+SCFace closed-set benchmark, about 3% above the vanilla backbone at a reported 2.1 GFLOPs.
Study and design of rot ational and translational with low illumination based face imag e recognition,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ForensicNet: Lightweight Attention-Enhanced MobileNetV2 for Automated Face Identification
MobileNetV2 + CBAM attention + two-phase layer unfreezing reaches 92.4% accuracy on a curated 68-identity LFW+SCFace closed-set benchmark, about 3% above the vanilla backbone at a reported 2.1 GFLOPs.