Learning-by-Novel-View-Synthesis for Full-Face Appearance-Based 3D Gaze Estimation

被引:0
|
作者
Qin, Jiawei [1 ]
Shimoyama, Takuru [1 ]
Sugano, Yusuke [1 ]
机构
[1] Univ Tokyo, Inst Ind Sci, Tokyo, Japan
关键词
CONSENSUS; TRACKING;
D O I
10.1109/CVPRW56347.2022.00546
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Despite recent advances in appearance-based gaze estimation techniques, the need for training data that covers the target head pose and gaze distribution remains a crucial challenge for practical deployment. This work examines a novel approach for synthesizing gaze estimation training data based on monocular 3D face reconstruction. Unlike prior works using multi-view reconstruction, photo-realistic CG models, or generative neural networks, our approach can manipulate and extend the head pose range of existing training data without any additional requirements. We introduce a projective matching procedure to align the reconstructed 3D facial mesh with the camera coordinate system and synthesize face images with accurate gaze labels. We also propose a mask-guided gaze estimation model and data augmentation strategies to further improve the estimation accuracy by taking advantage of synthetic training data. Experiments using multiple public datasets show that our approach significantly improves the estimation performance on challenging cross-dataset settings with non-overlapping gaze distributions.
引用
收藏
页码:4977 / 4987
页数:11
相关论文
共 50 条
  • [31] Multi-View Face Alignment Using 3D Shape Model for View Estimation
    Su, Yanchao
    Ai, Haizhou
    Lao, Shihong
    [J]. ADVANCES IN BIOMETRICS, 2009, 5558 : 179 - +
  • [32] A View Based Approach for Matching the 3D Appearance of Local Features
    Strachan, Euan
    Siebert, J. Paul
    [J]. IMAGE ANALYSIS AND RECOGNITION, 2013, 7950 : 364 - 372
  • [33] A 3D Gaze Estimation Method Based on Facial Feature Tracking
    Zhao, Xinbo
    Zou, Xiaochun
    Chi, Zheru
    [J]. 2012 INTERNATIONAL CONFERENCE ON COMPUTERIZED HEALTHCARE (ICCH), 2012, : 13 - 16
  • [34] 3D Face Recognition Based on Deep Learning
    Luo, Jing
    Hu, Fei
    Wang, Ruihuan
    [J]. 2019 IEEE INTERNATIONAL CONFERENCE ON MECHATRONICS AND AUTOMATION (ICMA), 2019, : 1576 - 1581
  • [35] Hand posture estimation by combining 2-D appearance-based and 3-D model-based approaches
    Shimada, N
    Kimura, K
    Shirai, Y
    Kuno, Y
    [J]. 15TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOL 3, PROCEEDINGS: IMAGE, SPEECH AND SIGNAL PROCESSING, 2000, : 705 - 708
  • [36] A Wide-View Parallax-Free Eye-Mark Recorder with a Hyperboloidal Half-Silvered Mirror and Appearance-Based Gaze Estimation
    Mori, Hiroki
    Sumiya, Erika
    Mashita, Tomohiro
    Kiyokawa, Kiyoshi
    Takemura, Haruo
    [J]. IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS, 2011, 17 (07) : 900 - 912
  • [37] A novel face recognition method based on 3D face model
    Liu Zhifang
    Wang Yunqiong
    You Zhisheng
    Zhao Minghua
    [J]. 2007 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND BIOMIMETICS, VOLS 1-5, 2007, : 439 - 444
  • [38] Appearance-Based Loop Detection from 3D Laser Data Using the Normal Distributions Transform
    Magnusson, Martin
    Andreasson, Henrik
    Nuechter, Andreas
    Lilienthal, Achim J.
    [J]. ICRA: 2009 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION, VOLS 1-7, 2009, : 3364 - +
  • [39] VIDEO-BASED FACE RECOGNITION BASED ON VIEW SYNTHESIS FROM 3D FACE MODEL RECONSTRUCTED FROM A SINGLE IMAGE
    Liao, Chia-Te
    Wang, Shu-Fang
    Lu, Yun-Jen
    Lai, Shang-Hong
    [J]. 2008 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO, VOLS 1-4, 2008, : 1589 - 1592
  • [40] Minimum variance estimation of 3D face shape from multi-view
    Zhang, Zhenqiu
    Hu, Yuxiao
    Yu, Tianli
    Huang, Thomas
    [J]. PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON AUTOMATIC FACE AND GESTURE RECOGNITION - PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE, 2006, : 547 - +