Optimized Lightweight Face Recognition Network Improves Performance on Embedded AI Platforms
Researchers have developed an improved MobileFaceNet model that balances accuracy and efficiency on embedded artificial intelligence platforms.
Researchers have developed an improved MobileFaceNet model that balances accuracy and efficiency on embedded artificial intelligence platforms.
A dual-stream deep learning architecture combining hierarchical taxonomic priors with attention-guided augmentation improved mosquito species classification precision to 84.6 percent ...
Researchers have combined convolutional neural networks and vision transformers with a three-loss training strategy to significantly improve content-based image retrieval ...
Researchers have built a compact, turkey-specific AI embedding that outperforms a much larger generic model at telling individual birds apart ...
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine