Synthesizing realistic profile faces is promising for more efficiently training deep pose-invariant models for large-scale unconstrained face recognition, by populating samples with extreme poses and avoiding tedious annotations. However, learning from synthetic faces may not achieve the desired performance due to the discrepancy between distributions of the synthetic and real face images. To narrow this gap, we propose a Dual-Agent Generative Adversarial Network (DA-GAN) model, which can improve the realism of a face simulator's output using unlabeled real faces, while preserving the identity information during the realism refinement. The dual agents are specifically designed for distinguishing real v.s.

This work is motivated by the engineering task of achieving a near state-of-the-art face recognition on a minimal computing budget running on an embedded system. Our main technical contribution centers around a novel training method, called Multibatch, for similarity learning, i.e., for the task of generating an invariant ``face signature'' through training pairs of ``same'' and ``not-same'' face images. The Multibatch method first generates signatures for a mini-batch of $k$ face images and then constructs an unbiased estimate of the full gradient by relying on all $k^2-k$ pairs from the mini-batch. We prove that the variance of the Multibatch estimator is bounded by $O(1/k^2)$, under some mild conditions. In contrast, the standard gradient estimator that relies on random $k/2$ pairs has a variance of order $1/k$. The smaller variance of the Multibatch estimator significantly speeds up the convergence rate of stochastic gradient descent. Using the Multibatch method we train a deep convolutional neural network that achieves an accuracy of $98.2\%$ on the LFW benchmark, while its prediction runtime takes only $30$msec on a single ARM Cortex A9 core. Furthermore, the entire training process took only 12 hours on a single Titan X GPU.

Among potential attacks against FR systems [Galbally et al., 2014, Biggio et al., 2015, Hadid et al., 2015, Mai et al., 2018, Marcel et al., 2023], the template inversion (TI) attack significantly jeopardizes the users' privacy. In a TI attack, the adversary gains access to templates stored in the FR system's database and aims

Synthetic face recognition (SFR) aims to generate synthetic face datasets that mimic the distribution of real face data, which allows for training face recognition models in a privacy-preserving manner.

Neural Information Processing Systems http://nips.cc/

Facial action units (AUs) recognition is essential for emotion analysis and has been widely applied in mental state analysis. Existing work on AU recognition usually requires big face dataset with accurate AU labels. However, manual AU annotation requires expertise and can be time-consuming. In this work, we propose asemi-supervised approach forAUrecognition utilizing alargenumber of web face images without AU labels and a small face dataset with AU labels inspired by the co-training methods.

Neural Information Processing Systems http://nips.cc/

Neural Information Processing Systems http://nips.cc/

Existing attacks almost perform on low-level instances, such as pixels and super-pixels, and rarely exploit semantic clues.

For the generator and the three discriminators, we use the FFHQ [2] and AAHQ datasets with 1024 1024 resolution. Hence, cooperating withGAN inversion methods, our framework is able to achieve arbitrary style transfer of a given face image. Wheni=0,allthelayersofthegenerator areinfluenced bythestylelatentcode. Result images of the directly concatenating method have similar face identities and head poses to their reference images, which means that this method leaks content information ofreference images to stylelatentcodes. However, for a reference image whose style is significantly different from that inAAHQ, ifdirectly feeding itinto BlendGAN, the style ofgenerated images maynotbesimilartothereference.