Real-Time Age and Gender Prediction
Real-Time Age and Gender Prediction
Volume 8, Issue 5, May 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Real-Time Age and Gender Prediction
Prof. Meera Sawalkar1
Department of Computer Engineering,
JSPM’s Rajarshi Shahu School of Engineering and Research, Pune, Maharashtra, India
Shubhangi Chavan2, Vaishnavi Jadhav3, Aakanksha Malusare4
BE Students, Computer Engineering, JSPM’s Rajarshi Shahu School
of Engineering and Research, Pune, Maharashtra, India.
Abstract:- Age and gender identification have grown to age and gender prediction in this paper. Wiki,
be important components of the biometric system, UTKFace, and Adience datasets were combined to
protection, and care. People frequently use it to get age- create a single collection of 18728 photos in order to
appropriate content. Social media makes use of it to increase the diversity of the training dataset. With this
broaden the reach of multilayer advertising and sizable mixed dataset, we were able to attain a real-time
promotions. Face detection has become so widely used accuracy of 48.5980.76. several experimental studies.
that we need to improve it using a variety of techniques
to get more accurate results. We have created a Paper Name: Use of Ensemble Modeling to Predict
lightweight deep Convolution neural network model for Age, Makeup, and Illumination from Faces.
age and generation prediction in this paper. Wiki, UTK
Face and Adience datasets have been combined to create Author: Kundan Nigam, Sahil Sharma
a single dataset of 18728 photos in order to increase the
Abstract: Classification based on factors like gender,
diversity of the training dataset. With this sizable mixed
appearance, age, and lighting has grown in importance
dataset, we were able to attain a real-time accuracy of
and potency for real-time applications. However, when
48.5980.76. several experimental studies.
compared in terms of optimization, the performance of
Keywords:- CNN(Convolution neural network), Age the current method is weak. To achieve more optimal
Prediction, Gender Prediction, Biometric System, Face findings, we used a variety of machine learning models
Detection. in our study article. We attempt to create ensemble
algorithms for reliable age classification ranges, which
I. INTRODUCTION are divided into four groups: child, young, middle, and
old. Similar to how makeup classification ranges are
Identification based on age and gender has become a split into partial makeup and over makeup, gender
crucial part of the biometric system, protection, and care. It classification ranges are split into male and female,
is widely used by people to get age-appropriate content. It is while illumination classification ranges are define.
used by social media to increase the reach of multilayer
promotions and advertising. We need to apply a variety of Paper Name: "Gender and Age based Census System
strategies to improve face detection because it is now being for Metropolitan Cities,"
used so frequently. In this paper, we develop a lightweight
deep convolution neural network model for predicting Author Shiv Mittal and Vikram Singh Rajput's
generation and age. To broaden the diversity of the training
dataset, the Wiki, UTK Face, and Adience datasets were Abstract: This study describes a clever computer-
pooled to provide a single dataset of 18728 images. vision system that determines a person's legal age and
gender, stores the information for census purposes, and
II. LITERATURE REVIEW aids in ensuring orderly entry of people into locations.
Deep learning-based facial image classifiers are used in
Paper Name: A Light weight Deep Convolutional the system to instantly determine a person's gender and
Neural Network Model for Real-Time Age and Gender age from an image of their face. The computer vision
Prediction. algorithm's prediction is used to generate the actuating
reaction, and it also maintains track of all faces it
Author: Md. Nahidul Islam Opu, Tanha Kabir Koly, encounters and the number of people who have those
Annesha Das and Ashim Dey. faces in common. You may monitor the log of the entire
surveillance process online.
Abstract: Age and gender identification have grown to
be important components of the biometric system, Paper Name: Real-Time CNN and ResNet Gender
protection, and care. People frequently use it to get age- Classification and Age Prediction.
appropriate content. Social media uses it to spread
multilayer advertising and promotions and broaden its Author: Asad Mustafa, Kevin Meehan
reach. Face detection has become so widely used that
we need to improve it using a variety of techniques to Abstract: The face is the most prominent element of
get more accurate results. We have created a the human body, and we can learn a lot about someone's
lightweight deep Convolution neural network model for identity from their facial traits, including age prediction
IJISRT23MAY1167 www.ijisrt.com 1227
Volume 8, Issue 5, May 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
and gen- der categorization. Computer vision (CV) is III. METHODOLOGY
currently being used to teach machines how to
understand and see the actual world. This study presents A. CNN ALGORITHM:
a unique artefact that uses a live stream from a camera Convolutional Neural Networks with a focus on image
source to recognize faces, categorize genders, and age- and video recognition applications. CNN is primarily
predict from human facial photos. The Keras CV library utilized for image analysis applications such segmentation,
and Convolutional Neural Networks (CNN) have both object detection, and picture recognition.
been utilized for training. This innovative study relies
on each features. Convolutional Neural Networks have three different
kinds of layers:
Paper Name: Human Age And Gender Classification Convolutional Layer: Each input neuron in a
using Convolutional Neural Network conventional neural network is connected to the following
Author: Mohammed Kamel Benkaddour hidden layer. Only a small portion of the input layer
Abstract:The development of intelligent systems neurons in CNN are connected to the hidden layer of
capable of effectively learning and recognising objects neurons.
is the major goal of the very active research fields of Pooling Layer: The pooling layer is used to make the
pattern recognition and automatic classification. The feature map less dimensional. Inside the CNN's hidden
usage of biometrics, which is generally utilised for layer, there will be numerous activation and pooling
security considerations, is integral to these applications. layers.
In the realm of research, the face modality has grown in Fully Connected Layer: Fully Connected Tiers make up
importance as a basic biometric technique. The the network's final few layers. The output from the last
objective of this work is to create a system for pooling or convolutional layer is passed into the fully
estimating age and gender from a facial image or real- connected layer, where it is flattened before being applied
time video using convolutional neural networks. In this
study, three CNN network models with various
architectures (number of filters, number of convolution
layers.
Fig. 1: CNN ALGORITHM
IV. GUI IMPLEMENTATION
A. WELCOME PAGE
Fig. 2: Welcome Page
IJISRT23MAY1167 www.ijisrt.com 1228
Volume 8, Issue 5, May 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
B. HOME PAGE
Fig. 3: Home Page
C. LOGIN PAGE:
Fig. 4: Login Page
D. REGISTRATION PAGE
Fig. 5: Registration Page
IJISRT23MAY1167 www.ijisrt.com 1229
Volume 8, Issue 5, May 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
V. RESULTS
Fig. 6: Result For Input 1
Fig. 7: Result For Input 2
VI. CONCLUSION allowing for reliable age and gender estimation. This project
holds promising potential for various applications such as
In conclusion, the real-time age and gender prediction targeted advertising, customer analysis, and human-
project using a Convolutional Neural Network (CNN) computer interaction systems. However, further refinement
demonstrates the ability to accurately predict the age and and evaluation are necessary to enhance the model's
gender of individuals in real-time. By leveraging the power robustness and generalizability across diverse populations
of deep learning and computer vision techniques, the CNN and environmental conditions.
model is capable of analyzing facial features and patterns,
IJISRT23MAY1167 www.ijisrt.com 1230
Volume 8, Issue 5, May 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
VII. FUTURE SCOPE
When changing a dataset, the same model can be
trained to predict the feelings of race etc. Age and gender
classifications can be used to predict age and gender in
uncontrolled real-time situations such as train stations,
banks, buses, airports, etc. For example, depending on the
number of male and female passengers by the age on the
train station, toilets and restrooms can be built to facilitate
transportation. Additionally, we'd like to improve the model
to account for unique situations like mask-covered faces.
REFERENCES
[1.] J. Brownlee, “How to Perform Face Detection with
Deep Learning,” Machine Learning Mastery, Jun. 02,
2019.
[2.] W. Ouyang et al., “DeepID-Net: Object Detection
with Deformable Part Based Convolutional Neural
Networks,” IEEE Trans. Pattern Anal. Mach. Intell.,
vol. 39, no. 7, pp. 1320–1334, Jul. 2017, doi:
10.1109/TPAMI.2016.2587642.
[3.] A. Voulodimos, N. Doulamis, A. Doulamis, and E.
Protopapadakis, “Deep Learning for Computer
Vision: A Brief Review,” Computational Intelligence
and Neuroscience, Feb. 01, 2018. (accessed Jul. 14,
2020).
[4.] Y.-L. Boureau, J. Ponce, and Y. LeCun, “A
Theoretical Analysis of Feature Pooling in Visual
Recognition,” p. 8, 2010.
[5.] S. Manasa, J. S. Abraham, A. Sharma, and K.
Himapoornashree, “Age, gender and emotion
detection using cnn,” International Journal of
Advanced Research in Computer Science, vol. 11, no.
Special Issue 1, p. 68, 2020.
[6.] O. Agbo-Ajala and S. Viriri, “Deeply learned
classifiers for age and gender predictions of unfiltered
faces,” The Scientific World Journal, vol. 2020, 2020.
[7.] A. Kharchevnikova and A. V. Savchenko, “Neural
networks in videobased age and gender recognition
on mobile platforms,” Optical Memory and Neural
Networks, vol. 27, no. 4, pp. 246–259, 2018.
[8.] M. Dileep and A. Danti, “Human age and gender
prediction based on neural networks and three sigma
control limits,” Applied Artificial Intelligence, vol.
32, no. 3, pp. 281–292, 2018.
[9.] Z. Zhang, Y. Song, and H. Qi, “Age
progression/regression by conditional adversarial
autoencoder,” in IEEE Conference on Computer
Vision and Pattern Recognition (CVPR). IEEE, 2017
[10.] C. B. Ng, Y. H. Tay, and B.-M. Goi, “Recognizing
Human Gender in Com- puterVision: A Survey,” in
PRICAI 2012: Trends in Artificial Intelligence,
Berlin, Heidelberg, 2012, pp. 335–346, doi:
10.1007/978-3- 642-32695-031.
