Disentangled Representation Learning GAN for Pose-Invariant Face Recognition
Citations
712 citations
560 citations
Cites methods from "Disentangled Representation Learnin..."
...To achieve this goal, current methods required carefully designed loss functions [26, 10, 32], introduction of additional attribute labels or features [23, 36, 3, 34, 31], or special architectures [11, 30] to train new models....
[...]
518 citations
Cites background from "Disentangled Representation Learnin..."
...Due to its ability to construct supplemental training samples, the GAN is very effective for small sample tasks, such as facial recognition [211] and complex noisy image denoising [31]....
[...]
509 citations
502 citations
Cites background from "Disentangled Representation Learnin..."
...While face recognition systems [42, 45] gain popularity, attackers present face spoofs (i....
[...]
References
111,197 citations
"Disentangled Representation Learnin..." refers methods in this paper
...Adam optimizer [15] is used with a learning rate of 0....
[...]
38,211 citations
"Disentangled Representation Learnin..." refers background in this paper
...[9] introduce GAN to learn generative models via an adversarial process....
[...]
...In practice, it is better for G to maximize log(D(G(z))) instead of minimizing log (1−D(G(z))) [9]....
[...]
...[9] suggest to alternate between k (usually k = 1) steps of optimizing D and one step of optimizing G....
[...]
...It is proved in [9] that this minimax game has a global optimum when the distribution pg of the synthetic samples and the distribution pd of the training samples are the same....
[...]
...GAN [9] can generate samples similar to a data distribution through a two-player game between a generator G and a discriminator D....
[...]
11,201 citations
8,289 citations
"Disentangled Representation Learnin..." refers background in this paper
...Recently, great progress is achieved with Deep Learning-based methods [28, 33]....
[...]
...[33] on LFW database, which consists of mostly nearfrontal faces....
[...]
...Second, other work focus on learning discriminative features directly from the non-frontal faces through either one joint model [28,33] or multiple pose-specific models [7,25]....
[...]
7,987 citations