Huang Yea-Shuan
ABSTRACT
Maintaining the identity while synthesizing the frontal view image is the most critical step in developing a "recognition via generation" framework. To this end, this paper investigates, tests and compares the performance of two deep learning architectures: Light-CNN and FaceNet. The Light-CNN is used to learn a robust feature for face verification tasks that produces a high-level facial identity accuracy over many traditional deep learning models. FaceNet, on the other hand, is a model to maps face images into a compact Euclidean space where distances directly represent a measure of face similarity. In our comparison, we use the TP-GAN model to perform several pre-processing stages. The face features are then extracted from the synthesized face images using Light-CNN and FaceNet as 256- and 128-dimensional representations, respectively. We evaluate the accuracy performances of Light-CNN and FaceNet architectures on Multi-PIE and FEI datasets.
- Junho Yim, Heechul Jung, ByungIn Yoo, Changkyu Choi, Dusik Park and Junmo Kim. 2015. Rotating your face using multi-task deep neural network. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 676–684. 10.1109/CVPR.2015.7298667Google Scholar
- Chao Xiong, Xiaowei Zhao, Danhang Tang, Karlekar Jayashree, Shuicheng Yan and Tae-Kyun Kim. 2015. Conditional convolutional neural network for modality-aware face recognition. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), Santiago, Chile, 3667 –3675. 10.1109/ICCV.2015.418Google ScholarDigital Library
- Xiangyu Zhu, Zhen Lei, Junjie Yan, Dong Yi and Stan Z-Li. 2015. High-fidelity Pose and Expression Normalization for face recognition in the wild. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, USA. 787–796. DOI: 10.1109/CVPR.2015.7298679Google Scholar
- Rui Huang, Shu Zhang, Tianyu Li and Ran He. 2017. Beyond face rotation: global and local perception GAN for photorealistic and identity preserving frontal view synthesis. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy. 2439–2448. 10.1109/ICCV.2017.267Google ScholarCross Ref
- Amir Ghodrati, Xu Jia, Marco Pedersoli and Tinne Tuytelaars. 2015. Towards Automatic Image Editing: Learning to See Another You. 1–9. https://arxiv.org/abs/1511.08446Google Scholar
- Shicun Chen, Yong Zhang, Baocai Yin and Boyue Wang. 2021. TRFH: Towards Real‑time Face Detection and Head Pose Estimation. Journal of Pattern Analysis and Applications. 1745–1755. https://link.springer.com/article/10.1007/s10044-021-01026-3Google ScholarDigital Library
- Xiaohu Shao, Junliang Xing, Ruihan Pan, Zhenghao Li, Xiangdong Zhou and Yu Shi. 2021. Multi-view Face Recognition Using Deep Attention-based Face Frontalization. In Proceedings of the IEEE International Conference on Multimedia and Expo (ICME), Shenzhen, China. 1–12. 10.1109/ICME51207.2021.9428396Google ScholarCross Ref
- Omar Elharrouss, Noor Almaadeed, Somaya Al-Maadeed and Fouad Khelifi. 2022. Pose-invariant Face Recognition with Multitask Cascade Networks. Journal of Neural Computing and Applications. 6039–6052. https://link.springer.com/article/10.1007/s00521-021-06690-4Google ScholarDigital Library
- Qiang Meng, Xiaqing Xu, Xiaobo Wang, Yang Qian, Yunxiao Qin, Zezheng Wang, Chenxu Zhao, Feng Zhou and Zhen Lei. 2021. PoseFace: Pose-Invariant Features and Pose-adaptive Loss for Face Recognition. 1–10. https://doi.org/10.48550/arXiv.2107.1172 1Google Scholar
- Pengyu Li, Biao Wang, Lei Zhang. 2021. Adversarial Pose Regression Network for Pose-invariant Face Recognitions. In Proceedings of the AAAI Conference on Artificial Intelligence. California, USA. 1940–1948. https://ojs.aaai.org/index.php/AAAI/article/view/16289Google ScholarCross Ref
- Kaipeng Zhang, Zhanpeng Zhang, Chia-Wen Cheng, Winston H. Hsu, Yu Qiao, Wei Liu and Tong Zhang. 2018. Super-identity Convolutional Neural Network for Face Hallucination. ECCV. 183–198. https://doi.org/10.48550/arXiv.1811.02328Google Scholar
- Xi Yin, Xiang Yu, Kihyuk Sohn, Xiaoming Liu and Manmohan Chandraker. 2017. Towards Large-pose Face Frontalization in the wild. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy. 1–10. https://doi.org/10.48550/arXiv.1704.06244Google ScholarCross Ref
- Augustus Odena, Christopher Olah and Jonathon Shlens. 2017. Conditional Image Synthesis with Auxiliary Classifier GANs. In Proceedings of the International Conference on Machine Learning. vol, 70. 2642–2651. https://proceedings.mlr.press/v70/odena17a.htmlGoogle ScholarDigital Library
- Bassel Zeno, Ilya kalinovskiy and Yuri Matveev. 2021. PFA-GAN: Pose Face Augmentation Based on Generative Adversarial Network. vol. 32, Is. 2. 425–440. https://doi.org/10.15388/21-INFOR443Google ScholarDigital Library
- Xiaoguang Tu, Jian Zhao, Qiankun Liu, Wenjie Ai, Guodong Guo, Zhifeng Li, Wei Liu and Jiashi Feng 2021. Joint Face Image Restoration and Frontalization for Recognition. IEEE Transactions. 1–14. https://arxiv.org/abs/2105.09907Google Scholar
- Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville and Yoshua Bengio. 2014. Generative adversarial networks. Advances in neural information processing systems.1–9. https://arxiv.org/abs/1406.2661?context=csGoogle Scholar
- Xiang Wu, Ran He, Zhenan Sun and Tieniu Tan. 2017. A light-CNN for deep face representation with noisy labels. IEEE Transactions. vol. 14, no. 8. 1–13. https://doi.org/10.48550/arXiv.1511.02683Google Scholar
- Florian Schroff, Dmitry Kalenichenko and James Philbin. 2015. Facenet: A unified embedding for face recognition and clustering. In Proceedings of the IEEE conference on computer vision and pattern recognition. Boston, MA, USA. 815–823. 10.1109/CVPR.2015.7298682Google ScholarCross Ref
- Xinzheng Xu, Meng Du, Huanxiu Guo, Jianying Chang and Xiaoyang Zhao. 2021. Lightweight FaceNet based on MobileNet. International Journal of Intelligence Science. vol. 11, no. 1. 1–15. 10.4236/ijis.2021.111001Google Scholar
- Ansh Abhay Balde, Abhay Jain and D. Patra. 2020. Improving facial recognition of FaceNet in a small dataset using DeepFakes. In Proceedings of the 4th International Conference on Computer, Communication and Signal Processing (ICCCSP). USA. 1–6. https://jglobal.jst.go.jp/en/detail?JGLOBAL_ID=202102237540769174Google Scholar
- Woong-Hee Lee, Mustafa Ozger, Ursula Challita and Ki Won Sung. 2021. Noise Learning Based Denoising Autoencoder. vol. 25, Is. 9. 2983–2987. 10.1109/LCOMM.2021.3091800Google Scholar
- Kaipeng Zhang, Zhanpeng Zhang, Zhifeng Li and Yu Qiao. 2016. Joint Face Detection and Alignment using Multi-Task Cascaded Convolutional Networks. In Proceeding of the 4th International Conference on Information Science and Control Engineering (ICISCE). Changsha, China. 1– 5. 10.1109/ICISCE.2017.95Google Scholar
- Ralph GrossaIain, Matthewsa Jeffrey, Cohnb Takeo Kanadea and Simon Bakerc. 2010. Multi-PIE. Image and vision computing. In Proceeding of the International Conference Auto Face Gesture Recognition. vol. 28, no. 5. 1 – 23. 10.1016/j.imavis.2009.08.002Google Scholar
- FEI. 2005-2006. Artificial Intelligence Laboratory of FEI. São Bernardo do Campo, São Paulo, Brazil. https://fei.edu.br/∼cet/facedatabase.html.Google Scholar
- Yanfei Liu and Junhua Chen. 2020. Unsupervised Face Frontalization for Pose-invariant Face Recognition. vol. 106. 1–9. https://doi.org/10.1016/j.imavis.2020.104093Google Scholar
Index Terms
- Comparative analysis of the Light-CNN and FaceNet methods for identifying and maintaining human faces
Recommendations
Robust Statistical Frontalization of Human and Animal Faces
The unconstrained acquisition of facial data in real-world conditions may result in face images with significant pose variations, illumination changes, and occlusions, affecting the performance of facial landmark localization and recognition methods. In ...
Identifying faces across variations in lighting: Psychophysics and computation
Humans have the ability to identify objects under varying lighting conditions with extraordinary accuracy. We investigated the behavioral aspects of this ability and compared it to the performance of the illumination cones (IC) model of Belhumeur and ...
Comments