Future of Information and Communication Conference (FICC) 2025
28-29 April 2025
Publication Links
IJACSA
Special Issues
Future of Information and Communication Conference (FICC)
Computing Conference
Intelligent Systems Conference (IntelliSys)
Future Technologies Conference (FTC)
International Journal of Advanced Computer Science and Applications(IJACSA), Volume 16 Issue 1, 2025.
Abstract: Face attribute estimation has several applications in computer vision, biometric systems, face verification /identification and image retrieval. The performance of face attribute estimation has been improved by using machine learning algorithms. In recent years, most algorithms have addressed this problem in multiple binary problem. Specifically, CNN-based approaches, which we can divide them into two classes; shared features and parts-based approaches. In shared features approach, the model uses two types of CNNs: one for feature extraction succeed by another one, for attribute classification. In the parts-based approaches, the approaches split the face image into multiple parts according to the geometric position of each attribute and train a CNN model for each part of the face. However, the shared features approach can handle attributes correlation but ignored attribute heterogeneity and gain in training time. On the other hand, the parts-based approaches can handle attributes heterogeneity but ignore attributes correlation and need more time in the training set compared with a shared feature approach. In this work, we propose a face attribute estimation method, which combined shared features and a parts-based approach into one model. Our model splits the input face image into five parts: whole image part, face part, face upper part, lower part, and nose part. In the same manner, the face attributes are subdivided into five groups according to the geometric position in the face image. We train shared feature model for each part, and we proposed an algorithm for feature selection task followed by AdaBoost algorithm to handle attribute classification task. Through a set of experiments using the LFWA and IIITM Face Emotion datasets, we demonstrate that our approach shows higher efficiency of face attribute estimation compared with the state-of-the art methods.
Abdelaali Benaiss, Otman Maarouf, Rachid El Ayachi, Mohamed Biniz and Mustapha Oujaoura, “Decoding Face Attributes: A Modified AlexNet Model with Emphasis on Correlation-Heterogeneity Relationship Between Facial Attributes” International Journal of Advanced Computer Science and Applications(IJACSA), 16(1), 2025. http://dx.doi.org/10.14569/IJACSA.2025.0160198
@article{Benaiss2025,
title = {Decoding Face Attributes: A Modified AlexNet Model with Emphasis on Correlation-Heterogeneity Relationship Between Facial Attributes},
journal = {International Journal of Advanced Computer Science and Applications},
doi = {10.14569/IJACSA.2025.0160198},
url = {http://dx.doi.org/10.14569/IJACSA.2025.0160198},
year = {2025},
publisher = {The Science and Information Organization},
volume = {16},
number = {1},
author = {Abdelaali Benaiss and Otman Maarouf and Rachid El Ayachi and Mohamed Biniz and Mustapha Oujaoura}
}
Copyright Statement: This is an open access article licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, even commercially as long as the original work is properly cited.