Renita Arianti Hapsari, Aji Purwinarko

Implementation of Convolutional Neural Network Algorithm Using Vgg-16 Architecture for Image Classification in Facial Images

Introduction

Implementation of convolutional neural network algorithm using vgg-16 architecture for image classification in facial images. Implement CNN with VGG-16 architecture for facial image classification, achieving 94% accuracy in gender recognition. Discover method impact and optimal epoch for high-performance results.

Abstract

Abstract: Face Recognition has now become a technology capable of recognizing facial patterns, facial image recognition is also used in various applications, for example in biological data recognition applications, digital image and video search, room security, and other applications. Purpose: This study aims to find out how the implementation of the CNN method with the VGG-16 architecture affects the classification of gender in facial images and how it affects the results. Methods/Study design/approach: In this study, we use the CNN method for data processing and build the program and use VGG-16 Architecture to build the model, then the tensorflow library for calling the required features such as when optimizing or for statistical plots and using the Confusion Matrix to determine the level of accuracy obtained. The desired results in this study are accuracy, precision, recall, and Fscore. Result/Findings: Classifying facial images using CNN with VGG-16 architecture provides an accuracy rate of 94%. From the results of this study it can be concluded that the model with the best accuracy is at epoch 20 compared to epoch 60, epoch 80, and epoch 100 which have previously been tested. Novelty/Originality/Value: The level of accuracy resulting from the implementation of the CNN method using the VGG-16 Architecture for image classification in facial images is quite good, resulting in an accuracy of 94%. Accuracy results were obtained from tests carried out by comparing several epoch values to produce the best accuracy of 94% using epoch 20.

Review

This paper presents an investigation into the implementation of a Convolutional Neural Network (CNN) algorithm using the VGG-16 architecture for image classification in facial images. The stated purpose is to specifically determine the impact of this method on the classification of gender. The authors utilize TensorFlow for program development and evaluation, employing a Confusion Matrix to derive key performance metrics such as accuracy, precision, recall, and F-score. A notable finding is an achieved accuracy rate of 94%, with the model demonstrating optimal performance at epoch 20 when compared to higher epoch values (60, 80, 100). The study highlights the practical relevance of facial recognition in various applications, from security to biological data analysis, making the chosen topic timely and pertinent. While the paper presents a clear objective and a promising accuracy result, several critical areas require more depth and clarity. A significant ambiguity exists regarding the classification task: the purpose mentions "classification of gender," but the results section broadly states "Classifying facial images" without explicitly confirming the 94% accuracy pertains specifically to gender classification. Crucially, the abstract lacks any details about the dataset used, including its size, source, characteristics (e.g., diversity, age groups, lighting conditions), and pre-processing steps. This omission severely limits the reproducibility of the study and the ability to assess the generalizability and potential biases of the reported performance. Furthermore, the "novelty/originality/value" section primarily reiterates the achieved accuracy and epoch comparison, without contextualizing the 94% accuracy against existing benchmarks or state-of-the-art methods in facial image or gender classification, making it difficult to ascertain the true contribution of this work. To enhance the scientific value and impact of this work, several improvements are recommended. Firstly, the authors should explicitly clarify whether the 94% accuracy refers to general facial image classification or specifically to gender classification. Detailed information about the dataset, including its source, size, class distribution, and any augmentation or pre-processing techniques, must be provided. It would also significantly strengthen the paper to include a comparative analysis of the achieved 94% accuracy against other established methods or pre-trained architectures for the same task, rather than just comparing different epoch values. Future work could explore the model's robustness to varying real-world conditions, investigate potential biases in gender classification (if applicable), and provide a more comprehensive discussion on the specific novel contributions beyond the direct implementation of a known architecture.

Full Text

Comments

You need to be logged in to post a comment.