Please use this identifier to cite or link to this item:
Title: Age classification using an optimized CNN architecture
Authors: Aydoğdu, M. Fatih
Demirci, Muhammed Fatih
Keywords: Age estimation
Convolutional neural networks
Deep learning
Issue Date: 2017
Publisher: Association for Computing Machinery
Source: 2017 International Conference on Compute and Data Analysis, ICCDA 2017, 19 May 2017 through 23 May 2017, , 130280
Abstract: With growing data size in multimedia systems, the need for successful image classification and retrieval systems becomes vital. Nevertheless, the performance of such systems is still limited for real-world applications. In this paper, we propose an optimized Convolutional Neural Network (CNN) architecture for the age classification problem. In order to justify the structure and depth of the proposed CNN-based framework, comprehensive experiments on a number of different CNN architectures are conducted. Based on the fitness of the age classification results with respect to success-error ratios, training times, and standard deviations of success rates; using exact, top-3 and 1-off criterion, the CNN architecture involving 4 convolutional layers and 2 fully connected layers is found to be superior to the other CNN-based architectures with different number of layers. We evaluate our method on a face database consisting of more than 55,000 images. © 2017 Association for Computing Machinery.
ISBN: 9781450352413
Appears in Collections:Bilgisayar Mühendisliği Bölümü / Department of Computer Engineering
Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection

Show full item record

CORE Recommender


checked on Sep 23, 2022

Page view(s)

checked on Feb 6, 2023

Google ScholarTM



Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.