Scribd
Upload
14 views4 pages

Lp5 DL Miniproject

Uploaded by

joellphns
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
14 views4 pages

Lp5 DL Miniproject

Uploaded by

joellphns
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

SCTRś Pune Institute of Computer Technology, Pune

A PROJECT REPORT ON
Gender and Age Detection using Deep Learning

SUBMITTED BY

Name: Joel Alphonso Roll No: 41333


Name: Chaitrali Ginimav Roll No: 41326
Name: Swaraj Gosavi Roll No: 41328

Under the guidance of


Prof. P.R Patil

DEPARTMENT OF COMPUTER ENGINEERING


1 Problem Statement
Implement a deep learning model that detects the gender (male/female) and estimates
the age group of a person from their facial image.

2 Learning Objectives
• Learn the basics of image preprocessing and dataset preparation.

• Understand how to use Convolutional Neural Networks (CNNs) for image classifi-
cation tasks.

• Apply transfer learning for gender and age classification.

3 Learning Outcomes
• Successfully build and train a CNN-based model for gender and age classification.

• Apply image preprocessing techniques such as face detection and resizing.

• Analyze and interpret the classification performance using metrics like accuracy
and loss.

4 Introduction
Gender and age detection has a wide range of applications in security, retail, and user ex-
perience personalization. With the rise of deep learning, Convolutional Neural Networks
have become the state-of-the-art technique for visual recognition tasks.

5 Objective
To design a deep learning model that takes facial images as input and predicts the cor-
responding gender and age group.

6 Dataset Description
The UTKFace dataset is used, which contains over 20,000 facial images with labels for
age, gender, and ethnicity.

• Format: Images in JPEG format

• Labels: Age, Gender

• Classes: Gender (0 = Male, 1 = Female), Age (0-116 years)

1
7 Methodology
7.1 Preprocessing
• Read images and extract labels from filenames.

• Resize images to 200x200 and normalize pixel values.

• Split into training, validation, and test sets.

7.2 Model Architecture


A CNN architecture with the following layers was used:

• Convolutional Layers + ReLU + Max Pooling

• Flatten + Dense Layers

• Output Layers: One for gender (binary classification), one for age (regression or
multi-class classification)

7.3 Training
• Optimizer: Adam

• Loss: Binary Crossentropy for gender, Categorical Crossentropy for age group

• Metrics: Accuracy, Loss

• Epochs: 20, Batch Size: 32

8 Experimental Setup
• Language: Python 3.9

• Framework: TensorFlow/Keras

• Hardware: NVIDIA GPU-enabled machine

• Libraries: OpenCV, NumPy, Matplotlib, Pandas

9 Results
The trained model achieved:

• Gender Classification Accuracy: ˜95%

• Age Group Prediction Accuracy: ˜80%

2
10 Conclusion
The deep learning model effectively classifies gender and predicts age groups with good
accuracy. Transfer learning and image preprocessing played a crucial role in achieving
high performance.

11 Future Work
• Improve age prediction with regression techniques or finer age groups.

• Apply face alignment techniques for better image normalization.

• Expand model to handle ethnicity and emotion recognition.

12 References
References
[1] Zhang, Zhifei, et al. ”Age Progression/Regression by Conditional Adversarial Autoen-
coder.” IEEE CVPR 2017. (UTKFace Dataset)

[2] Chollet, Francois. Deep Learning with Python. Manning Publications, 2018.

[3] OpenCV Library. https://opencv.org/

You might also like