Look Across Elapse: Disentangled Representation Learning and Photorealistic Cross-Age Face Synthesis for Age-Invariant Face Recognition

Jian Zhao, Yu Cheng, Yi Cheng, Yang Yang, Fang Zhao, Jianshu Li, Hengzhu Liu, Shuicheng Yan, Jiashi Feng
The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19), 2019.01

Despite the remarkable progress in face recognition related technologies, reliably recognizing faces across ages still remains a big challenge. The appearance of a human face changes substantially over time, resulting in significant intra-class variations. As opposed to current techniques for age-invariant face recognition, which either directly extract age-invariant features for recognition, or first synthesize a face that matches target age before feature extraction, we argue that it is more desirable to perform both tasks jointly so that they can leverage each other. To this end, we propose a deep Age-Invariant Model (AIM) for face recognition in the wild with three distinct novelties. First, AIM presents a novel unified deep architecture jointly performing cross-age face synthesis and recognition in a mutual boosting way. Second, AIM achieves continuous face rejuvenation/aging with remarkable photorealistic and identity-preserving properties, avoiding the requirement of paired data and the true age of testing samples. Third, we develop effective and novel training strategies for end-to-end learning the whole deep architecture, which generates powerful age-invariant face representations explicitly disentangled from the age variation. Moreover, we propose a new large-scale Cross-Age Face Recognition (CAFR) benchmark dataset to facilitate existing efforts and push the frontiers of age-invariant face recognition research. Extensive experiments on both our CAFR and several other cross-age datasets (MORPH, CACD and FG-NET) demonstrate the superiority of the proposed AIM model over the state-of-the-arts. Benchmarking our model on one of the most popular unconstrained face recognition datasets IJB-C additionally verifies the promising generalizability of AIM in recognizing faces in the wild.

Link: https://arxiv.org/abs/1809.00338

Back

Share

The ITS Summit Singapore 2017 @Marina Bay Sands

Panasonic R&D Center Singapore partook in a 2-day Intelligent Transportation Society (ITS) Summit Singapore 2017, which attracted more than 300 delegates, industry partners and speakers from the Asia Pacific countries. At this summit, we showcased our initiatives in autonomous automotive technologies and solutions as well as our efforts to accelerate Singapore’s Autonomous Vehicle landscape.

Paper accepted for publication in the IEEE Transactions on Systems, Man, and Cybernetics: Systems

Collaborating with Nanjing University of Science and Technology (NUST), Panasonic R&D Center Singapore co-authored a paper titled “Modular Lightweight Network…

Panasonic tops MS-Celeb-1M Challenges

Panasonic R&D Center Singapore, in collaboration with the National University of Singapore (NUS), has achieved the No. 1 position in MS-Celeb-1M Challenge 2017 , which is organized by Microsoft, for both Challenge 1 and Challenge 2. The results are in alignment with our mission to promote Panasonic’s outstanding Face Recognition technology worldwide.

IPBC Asia 2016 @Shanghai, China

PRDCSG-IPG attended the IPBC Asia (held in Shanghai on December 4-6, 2016), which is one of the leading events in the IP industry. IPBC Asia attracted top IP executives of global companies to gather and take center stage to discuss issues about the global IP development with a focus on how IP owners can strategically utilize their IP assets, especially in Asia.

Company News