Human Age Estimation from Facial Images Using Artificial Neural Network

Introduction
Face Images convey a significant amount of knowledge including information about identity, emotional state, ethnic origin, gender, age, and head orientation of a person shown in face image. This type of information plays a significant role during face-to-face communication between humans [1]. Above prospects of facial images can be used in emerging branch of Human Computer Interaction (HCI). Human age has following characteristics:
Aging is uncontrollable process: Aging cannot be delayed or advanced at will. It is slow and irreversible process.
Personal Age Patterns: The aging factor of a person is defined by his genetic structure as well as external factors like health, lifestyle, weather conditions, ethnicity, etc.
Aging Pattern is temporal data: Age and face patterns are vary with time. Age pattern at an instance affects all future patterns [2].
Thus, automatic age estimation, being an important technique in real world applications, has become difficult due to these characteristics. Not only these factors but sex of the person also plays a vital role in this process. For this process we need a collection sufficient data of images for training purpose which is partly eased due to public availability of aging database FG-NET which contains the necessary features in co-ordinate form and the age of that individual. The age range covered in this database is 0-69. Fortunately, a “complete” aging face database is unnecessary since human beings also learn to perceive facial ages from incomplete aging patterns [2]. We use multi-layered Artificial Neural Networks(ANN) for classification of the age group of a person. The result is classified into total eight groups of age-ranges.
The motivation for our work lies in various...

... middle of paper ...

...res", The 36th Intl. Conf. on Acoustics, Speech and Signal Processing, (ICASSP 2011), Prague, Czech Republic, May, 2011.
[13] Yinyin Liu, Janusz A. Starzyk, Zhen Zhu, “Optimizing Number Of Hidden Neurons in Neural Networks”, http://www.ohio.edu/people/starzykj/network/Research/Papers/Recent%20conferences../Hidden%20Neurons%20AIA2007_549-204.pdf.
[14] http://www.webpages.ttu.edu/dleverin/neural_network/neural_networks.html.
[15] S. Ziabakhsh, M. Payravi, H. Alavi-Rad and S. Ziabakhsh, "Color Image Compression Using Singular Value Decomposition", Proceedings of the first regional conference on new approaches in computer engineering and information technology, 2011.
[16] SaurabhKarsoliya, Approximating Number of Hidden layer neurons in Multiple Hidden Layer BPNN Architecture, International Journal of Engineering Trends and Technology- Volume3 Issue6- 2012, pp. 714-717.