A Unified Framework for Biphasic Facial Age Translation with Noisy-Semantic Guided Generative Adversarial Networks

TL;DR

This paper introduces a unified GAN-based framework for biphasic facial age translation that incorporates noisy-semantic guidance and disentangles facial features to produce more accurate age-progressed or regressed faces.

Contribution

The proposed framework uniquely integrates noisy-semantic layouts with dual sub-networks, ProjectionNet and ConstraintNet, employing attention and cycle-consistency for improved age translation fidelity.

Findings

Achieves state-of-the-art results on MORPH and CACD datasets.

Effectively models age-related facial changes with semantic supervision.

Demonstrates superior detail preservation in generated faces.

Abstract

Biphasic facial age translation aims at predicting the appearance of the input face at any age. Facial age translation has received considerable research attention in the last decade due to its practical value in cross-age face recognition and various entertainment applications. However, most existing methods model age changes between holistic images, regardless of the human face structure and the age-changing patterns of individual facial components. Consequently, the lack of semantic supervision will cause infidelity of generated faces in detail. To this end, we propose a unified framework for biphasic facial age translation with noisy-semantic guided generative adversarial networks. Structurally, we project the class-aware noisy semantic layouts to soft latent maps for the following injection operation on the individual facial parts. In particular, we introduce two sub-networks, ProjectionNet and ConstraintNet. ProjectionNet introduces the low-level structural semantic information with noise map and produces soft latent maps. ConstraintNet disentangles the high-level spatial features to constrain the soft latent maps, which endows more age-related context into the soft latent maps. Specifically, attention mechanism is employed in ConstraintNet for feature disentanglement. Meanwhile, in order to mine the strongest mapping ability of the network, we embed two types of learning strategies in the training procedure, supervised self-driven generation and unsupervised condition-driven cycle-consistent generation. As a result, extensive experiments conducted on MORPH and CACD datasets demonstrate the prominent ability of our proposed method which achieves state-of-the-art performance.