Development of Model for Estimation of Age Group of Drivers by using Soft Computing Tools

Author Name(s): Sanjay S Wankhede, Preeti Bajaj
Author Email: sanjay.wankhede@raisoni.net

Abstract

Road accidents are rapidly increased day to day, that affects the various humans’ life. Driver is the main parameter of the vehicle that sources the accidents. Many researchers interested to gather the data of the people who are accountable for the accidents. These papers mainly focus on the age group of the driver because age group is the most important factor for the driver’s behaviour. The age group estimation of the driver is based on his or her facial extraction mainly on the basis of multidimensional facial images. In this paper, we developed neural network model to analyze the age group of the driver using facial features extraction.

Keywords

Video, Frame, Face Detection, Neural Network, Age Estimation

Introduction

The road accidents occurs because driver behaviour, vehicle failure, roadway design and poor roadway maintenance. Over 90% of road traffic accidents involve human factor, there were major effects of age, experience and education on violation behaviour of traffic rules [1]. Those people who are really interested to collect the data for the people who are responsible for the accident or who are suffer from the accidents, this paper will really help them. Age group is also key parameter for the driver’s data collection. In this paper Age group of drivers is estimated. Automatic age estimation has rarely been explored. In a lot of behaviour of a human being the age development affects organization, size and look. Some characteristics of craniofacial morphology become visible only in human being of certain age range and transform during the aging. From birth to later life Changes in skin texture generally take place. According to Geng et al [2] Most of the change are cause by change in facial structures development like Cheeks are spread over a larger area, Chin becomes more prominent, reduces the free space on the surface of the skull. In addition to changes caused by facial structures development, some small changes in the skin take place, Facial hair turn out to be denser and change colour, Skin colour changes, Skin become darker, thinner, fewer flexible and more tough. Moreover, wrinkles, hang down cheeks and lowered bags under the eyes become visible, but most of the facial structures development occurs at an early age of the individual. [3]. And if we consider all face then all have dissimilar features for example some male have beard and mustache but some male didn’t have. Some male haveshort hair but some female also have short hair. There are many lacunas so that’s why we can’t justify the accurate age of the person from their face. For that many researcher are distribute their ages in some group such as child, young, youth and old. Or they do this work in the groups. In this paper 10 group is distributed as like 1-10, 11-15, 16-22, 23-27, 28-35, 3642, 43-50, 51-58, 59-65, and 67-80. No one can give the precise accuracy hence our aim is to minimize the age group, for that larger data base is required for training the neural network. There are many ways to estimate the age group of the drivers. Some researcher uses the landmark method. In which they use the landmark points on the face. In that some uses the 68 point or some use the 57 points on the face, is all up to the researchers. And in FGNET database they have some extracted feature. Some researcher manipulate the ratio method in that they utilize the ratio between the distance of the eyes and nose, and eye and mouth, or the ratio of the distance between mouth and nose and distance between the nose and eye and many more[4]. And some researcher calculates the distance between the eye and distance between the eye and nose. In that they have their individual reference parameters. And by using those parameters they calculate the age of person. But all they are facing with the same problem because of the insufficient data base, accurate age is not estimated.

Conclusion

Because of road accident increase day to day, which affects the human life, so this paper will help to reduce the road accident. Age group is important characteristic of the driver’s behavior, in this paper video of traffic collect two dimensional facial image. For estimation of age group of driver 263 parameters are calculated, that will help to control over the traffic violence.

214 total views, 1 views today

Download PDF File

About the author: admin