Dynamic Weight Allocation for Improved Age-Invariant Face Recognition and Age Estimation System: DMT-MFFCNN Approach

International Journal of Electronics and Communication Engineering

© 2023 by SSRG - IJECE Journal

Volume 10 Issue 7

Year of Publication : 2023

Authors : M. Rajababu, K. Srinivas, H. Ravisankar

10.14445/23488549/IJECE-V10I7P110

How to Cite?

M. Rajababu, K. Srinivas, H. Ravisankar, "Dynamic Weight Allocation for Improved Age-Invariant Face Recognition and Age Estimation System: DMT-MFFCNN Approach," SSRG International Journal of Electronics and Communication Engineering, vol. 10,  no. 7, pp. 97-107, 2023. Crossref, https://doi.org/10.14445/23488549/IJECE-V10I7P110

Abstract:

The investigation of Age Invariant Face Recognition (AIFR) and Age Estimation has received more attention in real-time applications. However, it is an exciting and challenging task due to the rapidly growing rate of image uploads on the internet in today's data-driven world. Multitask Learning (MTL) has shown great potential across various applications, outperforming single-task learning, mainly when applied in deep multitask networks. One of the most critical issues in MTL is determining the weight of each task. We propose a Dynamic Multitask and Multi-Feature Fusion CNN (DMT-MFFCNN) framework that can jointly perform face recognition and age estimation tasks by adopting the weights of the tasks dynamically based on the difficulty in training them. This approach avoids the exhausting and time-consuming process of manually tuning the weights. In particular, the proposed method can successfully enhance the multitask learning model's performance and can be quickly applied without hyperparameter tuning. We evaluated our model on the FGNET, CACD, and UTK Face datasets and a Live dataset AS-23 containing face images of different ages. In the proposed system, the dataset's images are first converted to grey level and Local Binary Pattern (LBP)images. Next, a Merged dataset is created by combining the features by fusion of original, grey level and LBP datasets. The experimentation is carried out with VGG-16, DenseNet-201 and ResNet-50 individual CNN models and the VDeRe-23 feature fusion CNN model. The final step is cross-validating the Merged dataset against a new unseen dataset. Among all the comparisons on Merged Dataset, the VGG-16 model achieved the best face recognition accuracy of 94.06%, and for age estimation with the same model, obtained an MAE value of 1.7 years. Additionally, with the feature fusion model (VDeRe-23), face recognition and age estimation tasks achieved an accuracy of 99.47% and an MAE of 1.67 years, respectively. The results of the extensive experiments have demonstrated that the proposed DMT-MFFCNN yields superior performance than state-of-the-art methods for both face recognition and age estimation tasks using single-task learning.

Keywords:

Age estimation, Age Invariant Face Recognition, Convolutional Neural Networks, Cross-dataset training, Merged datasets, Multitask Learning.

References:

[1] Ronan Collobert, and Jason Weston, “A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask Learning,” Proceedings of the 25th International Conference on Machine Learning, pp. 160-167, 2008.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Li Deng, Geoffrey Hinton, and Brian Kingsbury, “New Types of Deep Neural Network Learning for Speech Recognition and Related Applications: An Overview,” IEEE International Conference on Acoustics, Speech and Signal Processing, pp. 8599-8603, 2013.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Weihua Chen et al., “A Multi-Task Deep Network for Person Re-Identification,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 31, no. 1, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[4] S. Sandhya Kumari, and K. Sandhya Rani, “Selection of MSER Region based Ultrasound Doppler Scan Image Big Data Classification using a Faster RCNN Network,” International Journal of Computer Engineering in Research Trends, vol. 9, no. 10, pp. 184-192, 2022.
[CrossRef] [Publisher Link]
[5] Ross Girshick, “Fast R-CNN,” Proceedings of the IEEE International Conference on Computer Vision, pp. 1440–1448, 2015.
[Publisher Link]
[6] Hao Liu et al., “Ordinal Deep Learning for Facial Age Estimation,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 29, no. 2, pp. 486-501, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Yandong Wen, Zhifeng Li, and Yu Qiao, “Latent Factor Guided Convolutional Neural Networks for Age-Invariant Face Recognition,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4893–4901, 2016.
[CrossRef] [Google Scholar] [Publisher Link]
[8] Avishek Garain et al., “GRA_Net: A Deep Learning Model for Classification of Age and Gender from Facial Images,” IEEE Access, vol. 9, pp. 85672-85689, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[9] Muhammad Firdaus Mustapha et al., “Age Group Classification using Convolutional Neural Network (CNN),” Journal of Physics: Conference Series, vol. 2084, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[10] Beichen Zhang, and Yue Bao, “Cross-Dataset Learning for Age Estimation,” IEEE Access, vol. 10, pp. 24048-24055, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[11] Duc-Quang Vu et al., “Age and Gender Recognition using Multi-Task CNN,” Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, pp. 1937-1941, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Chenfei Xu, Qihe Liu, and Mao Ye, “Age Invariant Face Recognition and Retrieval by Coupled Auto-Encoder Networks,” Neurocomputing, vol. 222, pp. 62–71, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Yitong Wang et al., “Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition,” Proceedings of the European Conference on Computer Vision, vol. 11219, pp. 764–779, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[14] Rajeev Ranjan, Vishal M Patel, and Rama Chellappa, “Hyperface: A Deep Multi-Task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 41, no. 1, pp. 121-135, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[15] Sebastian Ruder, “An Overview of Multi-Task Learning in Deep Neural Networks,” Arxiv Preprint Arxiv:1706.05098, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[16] Yangjian Huang, and Haifeng Hu, “A Parallel Architecture of Age Adversarial Convolutional Neural Network for Cross-Age Face Recognition,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 31, no. 1, pp. 148-159, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[17] Alex Kendall, Yarin Gal, and Roberto Cipolla, “Multi-Task Learning using Uncertainty to Weigh Losses for Scene Geometry and Semantics,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7482–7491, 2018.
[Google Scholar] [Publisher Link]
[18] Chenggang Yan et al., “Age-Invariant Face Recognition by Multi-Feature Fusion and Decomposition with Self-Attention,” ACM Transactions on Multimedia Computing, Communications, and Applications, vol. 18, no. 15, pp. 1-18, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Krishna Dharavath, Amarnath Gaini, and Viod Adla, “A Parallel Deep Learning Approach for Age Invariant Face Recognition System,” International Conference on Computer Communication and Informatics, pp. 1-6, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[20] Yonglong Tian et al., “Pedestrian Detection Aided by Deep Learning Semantic Tasks,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5079-5087, 2015.
[Google Scholar] [Publisher Link]
[21] D. Betteena Sheryl Fernando et al., “Face Recognition for Home Security,” SSRG International Journal of Computer Science and Engineering, vol. 6, no. 10, pp. 7-12, 2019.
[CrossRef] [Publisher Link]
[22] Vinita Bhandiwad, “A Review Paper on 2D to 3D Face Recognition Technology,” SSRG International Journal of VLSI & Signal Processing, vol. 2, no. 2, pp. 10-13, 2015.
[CrossRef] [Publisher Link]
[23] Xi Yin, and Xiaoming Liu, “Multi-Task Convolutional Neural Network for Pose-Invariant Face Recognition,” IEEE Transactions on Image Processing, vol. 27, no. 2, pp. 964-975, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[24] Amirali Abdolrashidi et al., “Age and Gender Prediction from Face Images using Attentional Convolutional Network,” Arxiv preprint arXiv:2010.03791, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[25] Zhanpeng Zhang et al., “Learning Deep Representation for Face Alignment with Auxiliary Attributes,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 38, no. 5, pp. 918-930, 2015.
[CrossRef] [Google Scholar] [Publisher Link]
[26] Jia-Hong Lee et al., “Joint Estimation of Age and Gender from Unconstrained Face Images using Lightweight Multi-Task CNN for Mobile Applications,” IEEE Conference on Multimedia Information Processing and Retrieval, pp. 162-165, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[27] Jinbiao Yu, and Liping Jing, “A Joint Multi-Task CNN for Cross-Age Face Recognition,” 25th IEEE International Conference on Image Processing, pp. 2411-2415, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[28] Jie Fang et al., “Muti-Stage Learning for Gender and Age Prediction,” Neurocomputing, vol. 334, pp. 114-124, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[29] Se Hyun Nam et al., “Age Estimation by Super-Resolution Reconstruction Based on Adversarial Networks,” IEEE Access, vol. 8, pp. 17103-17120, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[30] Zhifei Zhang, Yang Song, and Hairong Qi, “Age Progression/Regression by Conditional Adversarial Autoencoder,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5810-5818, 2017.
[Google Scholar] [Publisher Link]
[31] Joyassree Sen et al., “Face Recognition using Deep Convolutional Network and One-shot Learning,” SSRG International Journal of Computer Science and Engineering, vol. 7, no. 4, pp. 23-29, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[32] T. Cootes, FG‐Net Face and Gesture Recognition Network, 2014.
[33] Bor-Chun Chen, Chu-Song Chen, and Winston H Hsu, “Cross-Age Reference Coding for Age-Invariant Face Recognition And Retrieval,” Computer Vision – ECCV, vol. 8694, pp. 768–783, 2014.
[CrossRef] [Google Scholar] [Publisher Link]
[34] Pooja G Nair, and R. Sneha, “A Review: Facial Recognition using Machine Learning,” International Journal of Recent Engineering Science, vol. 7, no. 3, pp. 85-89, 2020.
[Google Scholar] [Publisher Link]
[35] Zhenxing Niu et al., “Ordinal Regression with Multiple Output CNN for Age Estimation,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4920-4928, 2016.
[Google Scholar] [Publisher Link]
[36] Anju, and Anu Rani, “Face Recognition using PCA Gobor Filter SVM Techniques,” International Journal of Computer & Organization Trends, vol. 6, no. 3, pp. 20-23, 2016.
[Publisher Link]
[37] Opeyemi Oyelesi, and Akingbade Kayode Francis, “Face Recognition for Access Control using PCA Algorithm,” SSRG International Journal of VLSI & Signal Processing, vol. 4, no. 2, pp. 22-27, 2017.
[CrossRef] [Publisher Link]
[38] Karen Simonyan, and Andrew Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” Arxiv Preprint Arxiv:1409.1556, 2014.
[CrossRef] [Google Scholar] [Publisher Link]
[39] Kaiming He et al., “Deep Residual Learning for Image Recognition,” IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA, pp. 770–778, 2015.
[Google Scholar] [Publisher Link]
[40] Gao Huang et al., “Densely Connected Convolutional Networks,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, pp. 4700–4708, 2018.
[Google Scholar] [Publisher Link]
[41] Linkai Li et al., “Integrated Access Control System of Face Recognition and Non-Contact Temperature Measurement Based on Arduino,” International Journal of Computer and Organization Trends, vol. 12, no. 2, pp. 1-5, 2022.
[CrossRef] [Publisher Link]
[42] Jayamala Kumar Patil et al., “A Novel Method of Traffic Rule Compliance Detection using Morphological Operations at Toll Booths,” SSRG International Journal of Electrical and Electronics Engineering, vol. 10, no. 1, pp. 148-159, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[43] Nazia Begum, and Syed Noorullah, “Biometric Uncorrelated Face Recognition using Unsupervised Learning,” International Journal of Computer Engineering in Research Trends, vol. 7, no. 3, pp. 13-18, 2020.
[CrossRef] [Publisher Link]