A new algorithm for age recognition from facial images

  • Authors:
  • Mohammad Mahdi Dehshibi;Azam Bastanfard

  • Affiliations:
  • Department of Electrical, Computer and IT, Islamic Azad University, Qazvin Branch, Iran;Department of Computer Engineering, Islamic Azad University, Karaj Branch, Iran

  • Venue:
  • Signal Processing
  • Year:
  • 2010

Quantified Score

Hi-index 0.08

Visualization

Abstract

In this paper, a new algorithm for age-group recognition from frontal face image is presented. The algorithm classifies subjects into four different age categories in four key stages: Pre-processing, facial feature extraction by a novel geometric feature-based method, face feature analysis, and age classification. In order to apply the algorithm to the problem, a face image database focusing on people's age information is required. Because there were no such databases, we created a database for this purpose, which is called Iranian face database (IFDB). IFDB contains digital images of people from 1 to 85 years of age. After pre-processing, then primary features of the faces in the database will be accurately detected. Finally, a neural network is used to classify the face into age groups using computed facial feature ratios and wrinkle densities. Experimental results show that the algorithm identifies the age group with accuracy of 86.64%.