International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
FACIAL IMAGE BASED GENDER CLASSIFICATION SYSTEM USING DEEP LEARNING MODEL Dr. S. Nagarajan 1, R. Saravanamani2, T. Udhayaprakash3, K. Vijay4 1234Dept. of Computer Science and Engineering, Government College of Engineering, Srirangam, Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - This paper proposes a real-time gender
automatically learn discriminative features from facial images, eliminating the need for manual feature extraction. The system's architecture consists of multiple convolutional layers followed by fully connected layers for gender classification. Data preprocessing techniques such as normalization and augmentation are applied to enhance the model's generalization ability.
classification system leveraging deep learning and webcam input. Convolutional Neural Network (CNN) architecture is designed to extract hierarchical features from facial images. Training involves data preprocessing, model optimization, and validation. Integration with webcam feed is achieved using Open CV, allowing the real-time image acquisition and classification. Deployment involves web application development for seamless user interaction. Testing demonstrates robustness and accuracy, with evaluation metrics indicating reliable performance. Overall, the system offers a practical solution for real-time gender classification from facial images. The CNN architecture is crucial for extracting meaningful features from facial images. It typically consists of multiple convolutional and pooling layers followed by fully connected layers for classification. Data preprocessing steps may include normalization, resizing, and augmentation to enhance the model's ability to generalize. Model optimization techniques such as learning rate, scheduling and regularization help to improve the performance and prevent over fitting. Integration with OpenCV enables the system to capture live video streams from a webcam, process the frames, and perform gender classification in real time. This integration allows seamless interaction with the system, making it suitable for applications like video conferencing, surveillance, and human-computer interaction. Overall, this work demonstrates the feasibility and effectiveness of deep learning for real-time gender classification.
Face images contain valuable information for biometric recognition. Gender, age, and race can be extracted from faces, raising privacy concerns. GDPR regulates data usage, limiting information extraction without consent. Previous work used SAN to confound gender extraction from faces. We propose a gender detection system based on VGG19 for facial images, ensuring privacy. Our system integrates with OpenCV for real-time gender classification from webcam images. Experiments show improved gender detection while preserving face matching utility. This system addresses privacy concerns in biometric applications. Gender recognition from faces, a key human ability, is gaining interest for machine applications. It's used in social robotics and digital signage for personalized interactions and targeted advertising. Real-world applications require in reliable performance under challenging conditions, such as varying lighting and poses. Running gender recognition on embedded devices, like robots or cameras, is challenging due to limited resources. This paper proposes a deep learning architecture, based on Mobile Net v2, optimized for gender recognition. The network is designed to balance accuracy and speed, making it suitable for real-time applications on embedded devices with limited resources.
Key Words: Gender classification, Convolutional Neural Network, open CV, Deep learning, Transfer Learning.
1.INTRODUCTION
Crucial biometric task with applications in various fields, often studied using different modalities. Facial modality, being universal and acceptable, is widely used. Deep learning has greatly improved gender classification, surpassing traditional methods. However, the need for large training datasets raises privacy concerns when using real images. To address this, we investigate the use of fake data generated by GANs for training, focusing on privacy and data augmentation. This paper presents a new approach using Deep fake faces for training CNNs, evaluated on real datasets. Additionally, we offer a substantial dataset of over 200,000 deep fake faces to facilitate further research in this domain. The deep learning model is trained on a large dataset of labeled facial images to learn gender-specific features. Integration with Open CV enables the system to capture real-
Gender classification from facial images is a crucial task in computer vision with numerous applications such as security, marketing, and human-computer interaction. The goal is to automatically determine the gender of a person from an image and open CV. This study presents a facial image-based gender classification system that utilizes a deep learning model for efficient and accurate gender prediction. Deep learning models, such as convolutional neural networks (CNNs), are particularly well-suited for this task because they can automatically learn complex patterns and features from raw image data. These models can be trained on large datasets of labeled facial images to effectively distinguish between male and female faces. The proposed system leverages convolutional neural networks (CNNs) to
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 2283