0% found this document useful (0 votes)
135 views4 pages

Age and Gender Detection

This document summarizes a research paper on age and gender detection using computer vision and deep learning techniques. The paper outlines how computer vision works to analyze digital images and videos similarly to human vision. It then reviews previous literature on age and gender prediction using deep convolutional neural networks (CNNs). The paper describes the libraries used in the research, including OpenCV, TensorFlow, Keras, and Math. It provides a flowchart of the system workflow, which uses an OpenCV face detector model to identify faces in images and prepare them for analysis with a blobfromimage algorithm before predicting age and gender.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
135 views4 pages

Age and Gender Detection

This document summarizes a research paper on age and gender detection using computer vision and deep learning techniques. The paper outlines how computer vision works to analyze digital images and videos similarly to human vision. It then reviews previous literature on age and gender prediction using deep convolutional neural networks (CNNs). The paper describes the libraries used in the research, including OpenCV, TensorFlow, Keras, and Math. It provides a flowchart of the system workflow, which uses an OpenCV face detector model to identify faces in images and prepare them for analysis with a blobfromimage algorithm before predicting age and gender.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Age and Gender Detection


Parth Anand Wadkar Varad Wangikar Prof. Pramod Kanjalkar
Vishwakarma Institute of Vishwakarma Institute of Vishwakarma Institute of
Technology,Pune Technology,Pune Technology,Pune

Abstract:- Due to the technological advancements made II. COMPUTER VISION


by mankind, the 21st century represents the beginning of
an era of unfathomable accomplishments. The With the help of computer vision, computers are now
technologies indicated above can be used to our advantage able to analyse digital photos and movies as if they were real-
to determine someone's age and even gender based on world objects. The difficulties it encounters are mostly
their glance into a camera, photograph, or even a video. brought on by biological vision and the advantage humans
This study paper will carefully outline the entire process, have over machines due to the millions of years of
the approaches and algorithms that can be employed, the development our civilization has undergone. The computer
most accurate one, and how it all fits together. has difficulties with this because it is unable to comprehend
Additionally, it will emphasize how crucial it is and how biological vision. Computer vision primarily entails
applying it can improve our day-to-day lives. The main collecting, processing, analyzing, and comprehending digital
goal of the research is to develop a gender and age images in order to extract information from the real world and
detector using deep learning on the audience dataset that produce symbolic or numerical data that is used to make
can roughly predict the gender and age of a picture of decisions. This procedure involves various training, such as
someone's face. Additionally, to overcome the issues of object recognition tracking a video.
accuracy and timing in order to obtain the most accurate
predictions and results. Additionally, the map for how III. LITERATURE REVIEW:
this technology can be used for our advantage and a look
at the broad range of applications where it can be used, [1] B. Abirami , T.S. Subashini, V. Mahavaishnavi
from dating apps to matrimonial sites, from security ,“Gender and age prediction from real time facial images using
services, CCTV surveillance, and policing. CNN”, This research assessed how to automatically determine
a person's age and gender using real-time face photographs.
Keywords:- Convolution Neural Network, Artificial The haar cascade pretrained model was used for face
Intelligence, Machine Learning. identification, and the area of the found face was added to the
Caffenet CNN architecture.
I. Introduction
Inside this Caffenet model, a total of 1250 photos were
Images and videos are what keep today's world running, trained, with 650 pictures being female and 600 being male.
from security services to watching movies of adorable dogs, The system delivered a total accuracy of 68.89%.
but these images and videos can help bring about a change in
how people behave and can aid in the efficient and convenient [2] Sai Challa, Sowjany, Ruchitha Reddy, “Age and
operation of the authorities. Understanding the application of Gender Prediction using Face Recognition” In this paper,
the prediction model that will be created, how it may be authors demonstrate that by learning representations using
integrated into our daily lives, and how it can contribute to a deep-convolutional neural networks, they were able to
safer world. Our model will be able to determine a person's significantly improve age and gender recognition using the
age and gender simply by looking at an image or video, which Caffe Architectures of Deep Learning Framework (CNN).
could be CCTV surveillance footage, thanks to the prediction They suggest a considerably more straightforward
algorithm that will be used. But the thought occurs: Why is it convolutional net architecture that can be used even in the
required? I'll provide a straightforward illustration to address absence of learning data.
your query: A man on the run after stealing $5 million from
Pacific Standard Bank. Multiple fake identities have prevented [3] Vaibhav Malik,Vanshaj Rathi,Vanshika Jaiswal,
his identity from being proven up to this point. However, the “Compute Age and Gender of Person through ML”, Two key
police are aware of the person's age and gender. Everyone who inferences are drawn from their findings. First, CNN can be
fulfils the age and gender specifications will be extensively used to classify people's ages and genders more accurately,
examined using this algorithm, which can be used as a visual even when studying marked contemporaneous unconstrained
surveillance model in the surveillance cameras in a 1-mile samples that are substantially smaller in size. Second, by using
radius of the bank. our model, others working on similar projects or projects in
the future that will require more training data sets can further
enhance the system.

IJISRT22NOV1502 www.ijisrt.com 1697


Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IV. LIBRARISE USED V. FLOWCHART

• Opencv
OpenCV, which stands for "open source computer
vision," is a computer vision and machine learning library that
is free and open source. This library is primarily used for real-
time image and video processing, as well as for business
analytics. Additionally, it supports a variety of Deeo learning
frameworks, including Cafe and TenseFlow. Due to the
structure of Opencv, numerous shared or static libraries are
included in the package.

• Tensorflow
Google developed and made available the Python library
TensorFlow, which is used for quick numerical computation.
It is a foundation library and is primarily used to directly create
deep learning models, occasionally using wrapper libraries to
make the process that is built on top of tensorFlow more
straightforward.

• Keras
In essence, Keras is a high-level Application
Programming Interface (API) of Tensorflow; it offers the
fundamental building blocks and abstractions required for the
creation and delivery of machine learning solutions coupled
with a very high iteration rate. Without Keras, it would not
have been feasible to fully utilise Tensorflow's scalability or
its cross-platform capabilities. Keras supports engineers and
researchers and encourages users to do so. You may export the
Keras models to run in Google Chrome or any other browser
on any smartphone or mobile device, as well as on a sizable
cluster of GPUs or TPU.

• Math
Almost all programmes that require mathematical
functions use this library, which is the most basic one. Access Fig 1: System workflow
to them, as described by the C-standard, is made available by
this module. VI. FRAMEWORK

Complex numbers cannot be used with this module's Here, the preexisting model known as opencv face
functions since there is no support for them. Another module detector.pbtxt is used to construct the function highlightface(),
called "cmath" helps offer the support for all the complex which will receive the image from the data set or can take the
number functions that may be required in order to accomplish real-time image and detect the distinctive facial dimensions in
that. This distinction was made since the majority of the image. First, the model is used to identify the face's width
consumers don't need or comprehend complex numbers. As a and height. Second, the blobfromimage function, one of the
result, it is not included in the basic math module. Most users most essential deep learning algorithms, is introduced during
manipulate numbers directly to conduct both simple and image recognition (). The image is prepped using this
complex mathematical operations. Furthermore, receiving an function.
exception in an "if-else" or other conditional statement case is
simpler than receiving a complicated result and requiring the This technique works by attempting to elevate your
programmer to ascertain the result in order to execute the image to the same level of property as the source image. The
conditional statement. researchers that developed or used these models examined a
variety of photographs that they had amassed themselves and
stored in their own database. According to the level of
properties like brightness, profile, and saturation, those images
are totally different. In order to use the image for detection,
the function blobfomimage() strips the image of all variables
and additional factors, converting it into this minimal format.

IJISRT22NOV1502 www.ijisrt.com 1698


Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
VII. IMPLEMENTATION facehighlight() function is called with face detection
network configuration (facenet) and frame as parameters and
the output is stored in variables resultimg and faceboxes.

If the network does not recognise any face then


faceboxes list would be empty and hence a message is
displayed saying "no face detected".
These libraries, which were explained earlier in the
section on how to use libraries, have been imported. Imshow is used to display the result.

VIII. RESULT
The highlight face function that we created is this. In
order to successfully carry out this operation, the network and
the frame images are used as the parameters.

1. A copy of the framework is generated in order to face the


facial dimensions.
2. By using blobefromimages, the copies were transformed
into blobes.
3. After setting the input, the network receives a forward pass.
4. The confidence is defined in the range 0-1 for each value in
0 to 127 for the empty list face Boxes. Coordinates x1, y1, x2,
y2 are obtained when the confidence beats the threshold of
0.7(as defined in the parameter).
5. The coordinates are appended to the list faceBoxes.
6. Each list of coordinates in the faceBoxes accounts for the
rectangles on the image.
7. The function returns the copy of frame as well as list of
faceBoxes. Fig 01. Image test 1

VideoCapture() function from openCV library is used to


read the image. It either gets the image from argument parser
or else it will get a live image using the webcam i.e.
VideoCapture(0) by setting padding value as 20.

Now a while loop is used to continuously read the image


until any key is pressed. Live Stream from webcam is read and
then stored in hasframe and frame. A waitkey is again called
followed with break. This is when it is not a live video and an
image is read from argparse. Fig 02. Image test 2

IJISRT22NOV1502 www.ijisrt.com 1699


Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IX. CONCLUSION

In order to reliably identify the gender and age range of


a person from a single image of their face, deep learning and
convolutional neural networks will be used in this Prototype
moving forward. The image might be a part of a certain
dataset, it might be parsed once through the prototype, or it
might even be real-time if no arguments are parsed. The
expected age may fall within one of the following ranges, and
the predicted gender is binary, falling into the categories of
"Male" and "Female" (0-2). (4- 6). (8-12). (15-20), (21-24),
(25-32), (38-43), (48-53), (60-100).

REFERENCES

[1]. B.Abirami, T.S.Subashini, V.Mahavaishnavi,"Gender


and age prediction from real time facial images using
CNN",Selection and peer-review under responsibility of
the scientific committee of the International Conference
on Nanotechnology: Ideas, Innovation and Industries ,
Volume 33, Part 7, 2020, Pages 4708-4712
[2]. Vaibhav Malik1, Vanshaj Rathi2, Vanshika Jaiswal3,
Anil Kumar Dubey4, "Compute Age and Gender of
Person through ML" , International Research Journal of
Engineering and Technology (IRJET),Volume: 07 Issue:
06 | June 2020
[3]. Sai Teja Challa, Sowjanya Jindam, Ruchitha Reddy
Reddy, Kalathila Uthej , " Age and Gender Prediction
using Face Recognition , International Journal of
Engineering and Advanced Technology (IJEAT)
Volume-11 Issue-2, December 2021,
[4]. Use-case:https://www.cameralyze.co/blog/age-gender-
detection-top-use-
cases#:~:text=Age%20and%20gender%20detection%2
0is,but%20few%20solutions%20were%20available.

IJISRT22NOV1502 www.ijisrt.com 1700

You might also like