Research on face recognition algorithm based on improved convolution neural network

doi:10.1109/ICIEA.2018.8398186

Paper

Research on face recognition algorithm based on improved convolution neural network

Published May 1, 2018 · Liu Hui, Yu-jie Song

2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA)

UNKNOWN SJR score

16

Citations

0

Influential Citations

Full text Semantic Scholar

Abstract

Convolution neural network is a depth learning algorithm that can automatically extract features. A face recognition method based on convolution neural network and Fisher criterion is brought up to resolve the difficulty of poor property of convolution neural networks under small samples. First, a discriminant metric function is added in the cost function of the error and enhance the classification of the network. Then the face features by utilizing the modified convolution neural network are extracted. Finally, the advantage of support vector machine in the small sample, nonlinear and high dimension is used to classify the extracted features. The experimental results show that the face recognition algorithm which based on the Fisher neural network combined with SVM can achieve good results in the case of fewer samples.

Study Snapshot

Key takeawayThe face recognition algorithm based on Fisher neural network combined with SVM can achieve good results in fewer samples, improving face recognition performance.

PopulationOlder adults (50-71 years)

Sample size24

MethodsObservational

OutcomesBody Mass Index projections

ResultsSocial networks mitigate obesity in older groups.

Sign up to use Study Snapshot

Consensus is limited without an account. Create an account or sign in to get more searches and use the Study Snapshot.

Create an account

Paper

Research on face recognition algorithm based on improved convolution neural network

Sign up to use Study Snapshot

References

Deep learning for biological image classification

Deep learning techniques, particularly Convolutional Neural Networks, outperform traditional methods in predicting wood board quality, but the texture descriptor method shows competitive performance.

Voltage stability monitoring of power systems using reduced network and artificial neural network

This paper presents network reduction-based methodologies for monitoring voltage stability in power systems using artificial neural networks, enhancing estimation accuracy and robustness.

Local receptive field constrained deep networks

Local receptive field constrained deep networks (LRF-DNNs) improve face completion and reconstruction performance by using Gaussian receptive field constraints and a method for concurrently finding advantageous receptive field centers.

Face Recognition Based on Principal Component Analysis

Principal Component Analysis technique effectively recognizes faces in controlled environments, providing a cost-effective and efficient method for identifying individuals.

Face recognition method based on neural network

The face recognition method based on neural networks is highly robust to face image changes, achieving'similarity in spirit' and maintaining the same processing time for each face image during recognition.

3D face authentication and recognition based on bilateral symmetry analysis

Our novel 3D face authentication method, based on bilateral symmetry analysis, achieves a 10.8% equal error rate for variable facial expressions and 0.8% for normal expressions.

Citations

Smart Attendance Registration and Management System through AI

The proposed Split Invariant Artificial Neural Network (SIANN) enhances facial recognition speed and accuracy in smart attendance systems by incorporating deep learning techniques and specialized features.

Controllability-Constrained Deep Network Models for Enhanced Control of Dynamical Systems

This study proposes a control-theoretical method that enhances deep neural network models with controllability, leading to more efficient controllers and lower long-term prediction error in dynamic systems.

DCNN-based load balancing and computation offloading of Intelligent Video Surveillance with Edge Computing

The DIVS system, based on Deep Learning and edge computing, can save 68.2% to 72.4% of system utilisation compared to local execution, offering precise and low-latency video analysis.

Convolution Neural Network Approach for Facial Mask Detection

The YOLO v8 model effectively detects facial masks in images, ensuring compliance with health protocols and reducing false positives.

Improvement Airport Security System with Face Recognition Using Neural Network Based on the Arduino Uno Microcontroller

The proposed airport security system using an Arduino uno microcontroller and artificial neural network achieves 93.33% recognition accuracy and 2.67 seconds execution time, making it more reliable.