Note: This video may require joining the NVIDIA Developer Program or login
GTC Silicon Valley-2019 ID:S9566:Face Recognition: from Scientific Research to Commercial Products
Winston Hsu(National Taiwan University)
We'll discuss problems in face recognition such as face identification, verification, and attribute detection, and review the field's key technical components. This session will cover neural network design options, training strategies, and the physical meanings of the variant cost functions, as well as advanced issues involved in deploying face-recognition technologies, including scalability, lighting, training data, cross-domain learning and data augmentation strategies. In addiition, we'll discuss challenges involved in adapting face recognition from 2D to 3D signals and deploying face recognition in tangible products.