DADI INSTITUTE OF ENGINEERING & TECHNOLOGY

AGE DETECTION FROM FACES USING CNN

A
Proposed Project Power Point Presentation
Submitted in partial fulfilment of the requirements for the degree of
BACHELOR OF TECHNOLOGY
in
COMPUTER SCIENCE & ENGINEERING
By
D. Chandini (20U41A05C4)
K. Sumanth (20U41A0558)
K. Sai Varaha Mounika (20U41A0592)
M. Sai Yashwant (20U41A05B8)

Under the Supervision of

Mrs. M. Sharmila, Assistant Professor, CSE
Dadi Institute of Engineering & Technology

DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING

DADI INSTITUTE OF ENGINEERING & TECHNOLOGY
(AN AUTONOMOUS INSTITUTE)
ANAKAPALLE - 531002, VISAKHAPATNAM, A.P.
MARCH - 2024
CONTENTS

1. ABSTRACT
2. OBJECTIVE
3. INTRODUCTION
4. LITERATURE REVIEW
5. PROBLEM DEFINITION
6. SYSTEM STUDY
7. SYSTEM ANALYSIS
8. SYSTEM DESIGN
9. RESULTS
10. CONCLUSION
11. REFERENCES

D I E T
ABSTRACT

Age detection is a fundamental task in computer vision with various real-world applications,
including security systems, personalized user experiences, and healthcare. In this study, we propose a
Convolutional Neural Network (CNN) approach for age detection from facial images. Our model
utilizes deep learning techniques to extract discriminative features from facial images, allowing for
accurate age estimation. We employ a large-scale dataset containing diverse facial images across
different age groups to train and evaluate our model. Our results highlight the potential of deep
learning techniques, particularly CNNs, in addressing age detection tasks with practical implications
in various domains. We design and train a CNN model on a large-scale dataset of facial images
annotated with age labels. Overall, our proposed approach offers a promising solution for age
detection from facial images, with implications for various real world applications requiring accurate
and efficient age estimation.

Keywords: Age detection, Convolutional Neural Networks (CNNs), Deep learning, Facial wrinkles,
Image processing.
D I E T
OBJECTIVE
• The objective of age detection from faces using Convolutional Neural Networks (CNNs) typically
involves training a model to accurately predict the age of a person based on their facial features.
• This involves collecting a dataset of images labeled with corresponding ages, then training a CNN to
learn patterns and features that correlate with age.
• The goal is to optimize the model's performance so it can accurately predict ages from unseen facial
images.

Age detection from facial images has garnered significant interest in computer vision research due to its
wide range of applications in various domains such as security, entertainment, and healthcare. The
ability to estimate a person's age from their facial features has numerous practical implications,
including personalized content delivery, age-restricted access control, and demographic analysis.
Convolutional Neural Networks (CNNs) have emerged as a powerful tool for age estimation tasks,
leveraging their capacity to learn complex hierarchical features directly from raw image data.

D I E T
INTRODUCTION
• Age detection from facial images has garnered significant interest in computer vision
research due to its wide range of applications in various domains such as security,
entertainment, and healthcare. The ability to estimate a person's age from their facial
features has numerous practical implications, including personalized content delivery,
age-restricted access control, and demographic analysis.
• Convolutional Neural Networks (CNNs) have emerged as a powerful tool for age
estimation tasks, leveraging their capacity to learn complex hierarchical features
directly from raw image data. In this study, we explore the application of CNNs for age
detection from faces, aiming to develop an accurate and robust model capable of
predicting ages with high precision.
• This research not only has implications for practical applications but also provides
insights into the underlying mechanisms of facial aging and the potential for automated
age estimation in real-world scenarios.
D I E T
LITERATURE REVIEW
[1] Khaled EL Karazle, Valliappan Raman, Patrick Then (2022): Automatic age estimation from
facial images is an exciting machine learning topic that has attracted researchers’ attention over the
past several years. Numerous human–computer interaction applications, such as targeted marketing,
content access control, or soft-biometrics systems, employ age estimation models to carry out
secondary tasks such as user filtering or identification. Despite the vast array of applications that
could benefit from automatic age estimation, building an automatic age estimation system comes with
issues such as data disparity, the unique ageing pattern of each individual, and facial photo quality.

[2] P. Karthigayani, Dr. Sridhar Seshadri (2010): The proposed work is to find the age estimation
can be done using Morph data set using machine vision and Machine Learning. In the proposed work
Robot is used sensor the image in 3D format in bicubic transformation and that image can be
segmented using spatio temporal information for computation. Age estimation can be done using LBP
(local binary patterns). Multilayer neural network-based classifier is used to find the object
identification.

D I E T
PROBLEM DEFINITION
The problem of age detection from faces using CNNs involves building a model that can
accurately predict the age of individuals based on their facial images. Given a dataset of facial images
labeled with corresponding ages, the goal is to train a CNN to learn discriminative features and patterns
that correlate with age. The model should be capable of handling variations in facial expressions,
lighting conditions, and other factors that may affect age estimation accuracy. The objective is to
develop a robust and reliable system that can be deployed in various real-world applications, such as
age-based content recommendation, age verification systems, and demographic analysis. CNN is a deep
learning technique used in image and text recognition. CNN has proven to have remarkable
performance in understanding image segmentation, image classification problems, and detection. It was
found that the proposed model showed a better performance than pretrained models in terms of
performance evaluation criteria. The proposed model was found to produce better accuracy in the test
set, including 33,000 thousand images. As the CNN provides an efficient dense network which
performs the prediction or identification etc. efficiently. In the trained model we proposed age group
classification model we have better accuracy. We trained the model with a dataset of 33,000 thousand
raw images.
D I E T
SYSTEM STUDY

EXISTING MODEL:

1. “Deep Expectation of Apparent Age from a Single Image” (DEX) framework proposed by Rothe
et al. in 2018. This model leverages a pre-trained VGG-16 architecture and is trained on a large-scale
dataset of facial images annotated with apparent ages. The DEX framework achieves state-of-the-art
performance in age estimation tasks by effectively capturing facial features and learning robust
representations of age-related characteristics.

2. “Deeply Learned Deeply Learned” (DLDL) architecture introduced by Zhang et al. in 2017 is
another notable model for age estimation using CNNs. The DLDL model combines both local and
global features extracted from facial images, resulting in superior accuracy compared to traditional
methods. These models represent advancements in age detection from faces using CNNs and serve
as valuable benchmarks for future research in this area

D I E T
SYSTEM STUDY
Advantages:
• CNNs have shown remarkable accuracy in age detection tasks due to their ability to automatically
learn hierarchical features from facial images. This results in more precise age estimation
compared to traditional methods.
• CNNs can be optimized for real-time processing, allowing for fast and efficient age detection in
applications such as security systems and user interfaces.

Disadvantages:
• This model require large amounts of labeled data for training to achieve high accuracy. Limited or
biased datasets can lead to model performance degradation and potential biases in age predictions.
• This model are prone to overfitting, especially when trained on limited or noisy data. Overfitting
can result in poor generalization performance, where the model performs well on the training data
but fails to accurately predict ages on unseen data.

D I E T
SYSTEM STUDY
PROPOSED MODEL:
The main object of the chapter is about design processes of Age Detection from using faces using
CNN and theory of Age Detection. This chapter shows process of Age Detection. Then describes
simulation process of results. Age detection from facial images has garnered significant interest in
computer vision research due to its wide range of applications in various domains such as security,
entertainment, and healthcare. The ability to estimate a person's age from their facial features has
numerous practical implications, including personalized content delivery, age-restricted access
control, and demographic analysis. Convolutional Neural Networks (CNNs) have emerged as a
powerful tool for age estimation tasks, leveraging their capacity to learn complex hierarchical
features directly from raw image data. In this study, we explore the application of CNNs for age
detection from faces, aiming to develop an accurate and robust model capable of predicting ages with
high precision. The objective is to develop a robust and reliable system that can be deployed in
various real-world applications, such as age-based content recommendation, age verification systems,
and demographic analysis.

D I E T
SYSTEM ANALYSIS
SYSTEM PARAMETERS

• First age class 1–2 has subjects having

growing teeth, hairs, generally speaking
• For 3–9, the facial features start to grow
• For 10–20, subjects start to look more
mature
• For21–27 age class is of mature adults.
• Likewise for 28–45 age class46–65
would have more loud aging features
• And the last class, 66–116 have grey
hairs, aging skin, etc.

D I E T
SYSTEM ANALYSIS

SOFTWARE COMPONENTS:

Operating system : Windows 10+

Coding Language : Python3
Web-based user interface : HTML, CSS, JavaScript
Compiler : Jupyter Notebook and Anaconda

D I E T
ARCHITECTURE
Architecture:

D I E T
SYSTEM DESIGN

UML DIAGRAMS:

Use Case Diagram:

D I E T
SYSTEM DESIGN
Activity Diagram:

D I E T
RESULT

The result of age detection refers to the outcome produced by a machine learning model or
algorithm when given an input image, typically in the context of predicting the age of a person
depicted in the image. Overall, the result of age detection provides valuable insights into the
predicted age of individuals in images, along with measures of confidence and uncertainty, which
are essential for assessing the reliability and accuracy of the model's predictions. The primary result
of age detection is the predicted age of the individual in the input image. This prediction is
generated by the trained model based on the features extracted from the image during the training
process. The predicted age is usually represented as a numerical value.
D I E T
CONCLUSION
• Though many previous methods have addressed the problems of age detection, until recently,
much of this work has focused on constrained images taken in lab settings. Such settings do not
adequately reflect appearance variations common to the real-world images in social websites and
online repositories. The easy availability of huge image collections provides modern machine
learning based systems with effectively endless training data, though this data is not always
suitably labeled for supervised learning.
• Two important conclusions can be made from our results. First, CNN can be used to provide
improved age detection results, even considering the much smaller size of contemporary
unconstrained image sets labeled for age and gender. Second, the simplicity of our model implies
that more elaborate systems using more training data may well be capable of substantially
improving results beyond those reported here.
• Overall, I think the accuracy of the models is decent but can be improved further by using more
data, data augmentation and better network architectures. One can also try to use a regression
model instead of classification for Age detection if enough data is available.

D I E T
REFERENCES
1. A. Lanitis and C. J. Taylor, “Towards automatic face identification robust to ageing variation,”
in Proceedings of the 4th IEEE International Conference on Automatic Face and Gesture
Recognition, FG 2000, pp. 391–396, fra, March 2000.
2. Z. Li and A. K. Jain, “A discriminative model for age invariant face recognition,” IEEE
Transactions on Information Forensics and Security, vol. 6, no. 3, pp. 1028–1037, 2011.
3. N. Dayan, Skin aging handbook: An Integrated Approach to Biochemistry and Product
Development, Andrew William Press, 2008.
4. A. Amadasi, N. Merusi, and C. Cattaneo, “How reliable is apparent age at death on
cadavers?” International Journal of Legal Medicine, vol. 129, no. 4, pp. 1437–1596, 2015.
5. Eidinger, E.; Enbar, R.; Hassner, T. Age and Gender Estimation of Unfiltered Faces. IEEE Trans.
Inf. Forensics Secur. 2014, 9, 2170–2179.
6. Levi, G.; Hassncer, T. Age and gender classification using convolutional neural networks. In
Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition
Workshops (CVPRW), Boston, MA, USA, 7–12 June 2015; pp. 34–42.

D I E T
THANK YOU

D I E T