Age Progression and Regression with Spatial Attention Modules

Abstract

Age progression and regression refers to aesthetically rendering a given face image to present effects of face aging and rejuvenation, respectively. Although numerous studies have been conducted in this topic, there are two major problems: 1) multiple models are usually trained to simulate different age mappings, and 2) the photo-realism of generated face images is heavily influenced by the variation of training images in terms of pose, illumination, and background. To address these issues, in this paper, we propose a framework based on conditional Generative Adversarial Networks (cGANs) to achieve age progression and regression simultaneously. Particularly, since face aging and rejuvenation are largely different in terms of image translation patterns, we model these two processes using two separate generators, each dedicated to one age changing process. In addition, we exploit spatial attention mechanisms to limit image modifications to regions closely related to age changes, so that images with high visual fidelity could be synthesized for in-the-wild cases. Experiments on multiple datasets demonstrate the ability of our model in synthesizing lifelike face images at desired ages with personalized features well preserved, and keeping age-irrelevant regions unchanged.

AgeFlow: Conditional Age Progression and Regression with Normalizing Flows

“…As illustrated in Figure 2, the proposed invertible conditional translation module (ICTM) T leverages the invertible network to achieve both age progression and regression, where the cycle consistency can be easily achieved due to its reversibility. In a sense, the inverse function of T plays a similar role in [Song et al, 2018;Li et al, 2020b] that achieves age regression independent of age progression. To further keep the flexibility of the cGANs-based method that achieves age progression/regression with only changing the target age condition, the translation module concatenates the latent vector z s with the distributions of target age condition estimated from the prior generator to achieve conditional translation.…”

“…However, the translation G : I s → I t is highly under-constrained due to the unpaired training of GANs, misleading the model to learn the patterns other than face aging/rejuvenation. One possible solution is to regularize the mappings with cycleconsistency [Zhu et al, 2017], which learns backwards translation I t → I s using a separate model [Song et al, 2018;Li et al, 2020b] to approximately reconstruct I s . However, the cycle-consistency brings more uncertainty of age mappings and cannot guarantee the bijective age mapping, especially when the gap between c s and c t becomes large.…”

“…We randomly divided the dataset into two parts without identities overlapping: 80% for training and the remaining for testing. Following the mainstream works [Yang et al, 2018;Liu et al, 2019;Li et al, 2020b;Huang et al, 2021a], we divided the face images into n = 4 age groups, i.e., 30-, 31-40, 41-50, 50+.…”

“…Initial attempts [Yang et al, 2018;Liu et al, 2019;Huang et al, 2021a] mainly focus on training several independent models for each age mapping. Considering the different patterns of age progression and regression, [Song et al, 2018] used two separate generators to model both age changing processes separately, and [Li et al, 2020b] further extended it with a spatial attention mechanism. However, they usually suffer from considerably computational costs and are limited with the age label as prior.…”

Deep Face Age Progression: A Survey

“…Wang et al [19] 2018 Age group transitions CACD First to include both identity-preservation and age classification loss. Zeng et al [82] 2018 Age group transitions UTKFace, FG-NET Zhu et al [71] 2018 Year-accurate ageing MORPH-II, UTKFace, FG-NET Gou et al [83] 2019 Age group transitions Webcrawled DB (private) Li et al [84] 2019 Age group transitions MORPH-II, CACD, UTKFace Dual cGAN (see [81]) with spatial attention mechanism. Liu et al [85] 2019 Age group transitions MORPH-II, CACD, IMDb-Wiki cGAN conditioned on age and gender.…”