e-ISSN: 2582-5208

International Research Journal of Modernization in Engineering Technology and Science

( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:03/Issue:07/July-2021 Impact Factor- 5.354 www.irjmets.com

AGE GENDER AND EMOTION IDENTIFICATION USING FACE RECOGNITION

Harshitha H S*1, Soujanya CK*2, Suhas S K*3, Swathi S Gowda*4, Dr. Shashikala S V*5
*1,2,3,4 UG Student, Department of Information Science &Engineering, Bgsit, Mandya,
Karnataka, India.
*5Faculty, Professor Department of Computer Science & Engineering, Bgsit, Mandya,
Karnataka, India.
ABSTRACT
Face detection and recognition from a photo or video is nowadays a very wall-liked topic. Age, gender and
emotion recognition is done by face detection using we cameras, we designed real time face recognition based
on IP camera and image set algorithm by way of OpenCV and Python programming development. The system
has three parts detection part, training part and recognition part. we have used HAAR cascading algorithm
applied to image classification area. Facial expression can give some future references in the human machine
interaction. Human facial image processing has been a very popular and obsessed topic research issues for
years now. The capabilities of age and gender identification using images was quite effective. Personal
monitoring and recognition are also important in age, gender and emotion recognition.
Keywords: Face recognition, emotion detection, OpenCV, NumPy, HAAR cascade.
I. INTRODUCTION
Face detection and recognition is technology which is used to identify a person’s age, gender and emotion from
an image or web camera. Nowadays Human behaviors sometimes it will be uncertain so identifying human
emotions and behavior from recorded video and image or using web camera is important nowadays, so Age
and gender identification using image is also prominent. these days we are using OpenCV and NumPy. OpenCV
will supports vast programming languages those are C++, python, java, etc., and it is available in other stands
that will includes so many operating systems like Linux, OS X, Android, Windows and iOS. Interfaces for high-
speed GPU operations based on CUDA and OpenCL are also under active development. OpenCV-python is
basically a python application programming interface for the OpenCV and here we are combining the best
qualities of OpenCV C++ application programming interface and the python language. And here we also using
HAAR CASCADE algorithm to our project.
HAAR cascade algorithm is machine learning object-oriented algorithm that is used for identification of objects
in image or photo or video based on the lots of positive and negative images. Here positive images these images
contain the images that we want the classifier to identify, negative images these are other images that will not
contains the objects that we are going to detect. The HAAR cascading algorithm have four major steps HAAR
feature selection, creating integral images, ad boost training, cascading classifiers these will detect the face and
body parts in images but can be trained to identify any object. Now let’s take face detection as a example
initially, the HAAR cascade algorithm needs a lot of positive images of a face and negative images without face
to train the classifier. and after that we need extract the feature from it. Firstly we need to collect the HAAR
features after that adjacent rectangular regions at a specific location in a detection window, sums up the pixels
intensities in each of the region and calculates the difference between the sums.
II. METHODOLOGY
Initial steps require the input which is already stored in system as a dataset of some images. After that
retrieving image from an external source for further processing takes place which is also called image
acquisition. To know image has been successfully captured and saved into memory following visualization.
Next is image filtering which is used to curb the high frequency in the image for filtering image it uses gaussian
algorithm. It is followed by preprocessing for smoothing of an image, thresholding and also edge detection of
the images or photos. It further follows the features extraction which starts from a first set of measured data
and builds derive values intended to be informative and non-redundant also for character recognition as it
defines the behavior of images. The very next stage is segmentation it is the process of partitioning a digital
image into multiple segments. The goal is to simply changes the representation of an image into something that
www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science
[1553]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:03/Issue:07/July-2021 Impact Factor- 5.354 www.irjmets.com
is more meaning full and easier to analyze and convert them into pixel formats. Next is train or test dataset here
in this stage we do training and testing of data set which we have given image. If we do 80% of training and
20% of testing this is the ratio of training and testing. if we want, we can also do validation. It uses HAAR
cascading algorithm which helps in object recognition and also in identifying the age and gender of the image.
HAAR cascade algorithm is a machine learning based topic where it contains so many positive and negative
images. After following all these above-mentioned steps, we get the output. Since approach to classify gender
and age from image of human faces which is an essential part of our method.
III. RESULTS AND DISCUSSION
The Datasets are first trained for face detection (skin texture, color, jawline, facial hair, chin, nose), Feature
extraction then it classifies the result initially we train the datasets, then it detects the face using webcam and
extracts the features of taken image and classifies them. the face detection process is to extract the face area
from the background of input images with various lighting conditions in object detection feature extraction
place an important role. Here, it includes shapes, movement, color, the texture of the facial image. It extracts
meaning full information from an image compared to the original image. Feature extraction greatly reduces
information of image, which has advantages in storage. The feature classification state recognizes facial images
and group them according to certain classes and help them reduce dimensions help to remove the non-
essential information from input images to recognize the face. Nowadays age and gender identification are a
very popular topic it is useful for now facial features play very significant role because of social interaction like
access control and all now we are taking about other work that they can increase the acquires of the
identification. here we are using Haar cascading algorithm this field is a very important in computer science
hence the application now depends on identification of gender, age, and emotion of human it has to be precise
and perfect.
In age estimation we get the gender identification and then in age we get age identification and in image
recognition we got some of the emotions, which results in achieving three goals: age, gender and emotion
recognition using CNN.
Gender and age classification is the facial features that play a significant role in social interaction. Age and
gender classification in face images play important tasks in online activities, for example access control, visual
observation and so on.
Emotion detection on face images can communicate and impact an individual’s feelings. It is one’s curiosity or
interest and to basic understating of the feelings passed on by the images and how the visual content of the
image implies emotions (i.e., happy, sad, angry, fear and neutral).

Figure :1 System Analysis

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[1554]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:03/Issue:07/July-2021 Impact Factor- 5.354 www.irjmets.com
Training of datasets are mainly important in here we have train the certain amount of images after the training
we have do the face detection here we are going to face our image or video or face then we will have to extract
the features after the feature extraction then after that we will do the classification after classifying the images
we have do the learning model part then here prediction of labels after the labels predictions CNN
classifications then here it will go to emotion recognition, age recognition and gender recognition in emotion
recognition we get happy, sad, neutral, angry and sad and in age we will get the age and after that we get gender
estimation in the process.

Figure :2 Shows how those features applicable

IV. CONCLUSION
In this paper we developed a combined system that is for face recognition it will using by OpenCV and NumPy.
These are used for the detection of age, gender and emotion from images. The images of a persons which are
basically trained and tested before recognition. For better recognition and detection, a bit other features can be
improved further. In the future the technologies will be more and more advanced features will be added to the
system of recognition.
V. REFERENCES
[1] P. Ekman, "Emotional and conversational nonverbal signals," in Language, knowledge, and
representation: Springer, pp. 39-50.
[2] K. Sandeep and A. Rajagopalan, "Human Face Detection in Cluttered Color Images Using Skin Color,
Edge Information," in ICVGIP,
[3] M. Zabłocki, K. Go, D. Frejlichowski, and R. Hofman, “Intelligent video surveillance systems for public
spaces – a survey,” J. Theor. Appl. Comput. Sci., vol. 8, no. 4, pp. 13–27, 2014
[4] T. A. McGregor, R. L. Klatzky, C. Hamilton, and S. J. Lederman, "Haptic classification of facial identity in
2D displays: Configural versus feature-based processing," Haptics, IEEE Transactions on, vol. 3, no. 1,
pp. 48-55, 2010.
[5] M. Hussain, A. Al-Haiqi, A. A. Zaidan, B. B. Zaidan, M. L. M. Kiah, N. B. Anuar, and M. Abdulnabi, “The
landscape of research on smartphone medical apps: Coherent taxonomy, motivations, open challenges
and recommendations,” Comput. Methods Programs Biomed., vol. 122, no. 3, pp. 393–408, 2015.
[6] D. Tran, L. Bourdev, R. Fergus, L. Torresani, and M. Paluri, “Learning spatiotemporal features with 3D
convolutional networks,” Proc. IEEE Int. Conf. Comput. Vis., vol. 11–18–Dece, pp. 4489–4497, 2016
[7] B. Sielly, J. Costa, C. G. Bezerra, L. A. Guedes, P. P. Angelov, and N. Z. Norte, “Unsupervised Classification
of Data Streams based on Typicality and Eccentricity Data Analytics,” in EEE International Conference
on Fuzzy Systems (FUZZ-IEEE), 2016, pp. 58–63.
[8] R. Sharma, T. S. Ashwin, and R. M. R. Guddeti, “A Novel Real-Time Face Detection System Using
Modified Affine Transformation and Haar Cascades,” in Recent Findings in Intelligent Computing
Techniques, Springer, 2019, pp. 193–204.

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[1555]