scholarly journals Two sources of bias explain errors in facial age estimation

2018 ◽  
Vol 5 (10) ◽  
pp. 180841 ◽  
Author(s):  
Colin W. G. Clifford ◽  
Tamara L. Watson ◽  
David White
Keyword(s):  
Social Interactions ◽  
Age Estimation ◽  
Estimation Error ◽  
Age Groups ◽  
Large Database ◽  
Police Investigations ◽  
Immediate Experience ◽  
Improving Accuracy ◽  
Facial Age ◽  
Age Estimates

Accurate age estimates underpin our everyday social interactions, the provision of age-restricted services and police investigations. Previous work suggests that these judgements are error-prone, but the processes giving rise to these errors are not understood. Here, we present the first systematic test of bias in age estimation using a large database of standardized passport images of heterogeneous ages ( n = 3948). In three experiments, we tested a range of perceiver age groups ( n = 84), and found average age estimation error to be approximately 8 years. We show that this error can be attributed to two separable sources of bias. First, and accounting for the vast majority of variance, our results show an assimilative serial dependency whereby estimates are systematically biased towards the age of the preceding face. Second, younger faces are generally perceived to be older than they are, and older faces to be younger. In combination, these biases account for around 95% of variance in age estimates. We conclude that perception of age is modulated by representations that encode both a viewer's recent and normative exposure to faces. The finding that age perception is subject to strong top-down influences based on our immediate experience has implications for our understanding of perceptual processes involved in face perception, and for improving accuracy of age estimation in important real-world tasks.

Accepting final counts from repeat readings of otoliths: should a common criterion apply to the age estimation of fish?

2010 ◽  
Vol 61 (10) ◽  
pp. 1171 ◽  
Author(s):  
Ross J. Marriott ◽  
Bruce D. Mapstone ◽  
Aaron C. Ballagh ◽  
Leanne M. Currey ◽  
Ann Penny ◽  
...  
Keyword(s):  
Count Data ◽  
Age Estimation ◽  
Age Groups ◽  
Multiple Criteria ◽  
Older Age ◽  
Explanatory Model ◽  
Probability Of Error ◽  
Exact Agreement ◽  
Common Criterion ◽  
Age Estimates

Multiple readings of otoliths are often carried out to assess the repeatability and reliability of increment counts for estimating fish age. Various criteria have been used to assign or discard age estimates from repeated counts when discrepancies occur although the reasons for doing so are usually not stated or justified. Trends in relative frequencies (percentage disagreement, PD) and magnitudes (inter-read discrepancy, IRD) of otolith-count discrepancies were explored for 15 species of fish collected from a range of locations around Australia to explore generality in the best explanatory model(s) for otolith-count discrepancies and, hence, the most appropriate criterion for accepting or rejecting age estimates from multiple-count data. Increasing discrepancies with increasing age, according to a constant per-increment probability of error, was the best-approximating model for 9 of the 15 species for PD data but for only two species for IRD data. Our results indicated disproportionately higher rates of rejection of estimates from older age groups if exact agreement between repeated counts was required for age acceptance. Results varied with the reader, region and the method of otolith reading, indicating that multiple criteria for accepting or rejecting counts from multiple readings may be required among or even within species.

scholarly journals Facial age estimation using SFTA and deep neural network

2018 ◽  
Vol 7 (2) ◽  
pp. 281
Author(s):  
Deepa Nagarajan ◽  
T. Sasipraba
Keyword(s):  
Neural Network ◽  
Feature Extraction ◽  
Age Estimation ◽  
Deep Neural Network ◽  
Age Groups ◽  
Histogram Equalization ◽  
Age Group ◽  
Final Layer ◽  
Feature Values ◽  
Facial Age

This paper construes the toils in facial age estimation in images. The fact that manual age estimation is indeed hard rising out the urge for digital age estimation. To make estimation precise many works have been carried out by considering a lot of constraints. In this paper, facial age estimation is done more accurately. SFTA method is used for feature extraction and meticulous results are obtained for all age groups. Histogram equalization is done using the Otsu algorithm and three layered Deep Neural Network is used to classify the age group. In a Deep neural network, softmax normalization is done in the final layer to preserve the outlier values. By extracting 45 feature values concerning color and gradient, key point descriptor, orientation, shape and texture better estimation are obtained.

A Review of Different Techniques of Age Estimation from Human Face

2018 ◽  
Vol 4 (10) ◽  
pp. 6
Author(s):  
Khemchandra Patel ◽  
Dr. Kamlesh Namdev
Keyword(s):  
Age Estimation ◽  
Real Life ◽  
Extraction Methods ◽  
Human Face ◽  
Age Changes ◽  
Missing Children ◽  
Different Types ◽  
Age Variation ◽  
Human Faces ◽  
Facial Age

Age changes cause major variations in the appearance of human faces. Due to many lifestyle factors, it is difficult to precisely predict how individuals may look with advancing years or how they looked with "retreating" years. This paper is a review of age variation methods and techniques, which is useful to capture wanted fugitives, finding missing children, updating employee databases, enhance powerful visual effect in film, television, gaming field. Currently there are many different methods available for age variation. Each has their own advantages and purpose. Because of its real life applications, researchers have shown great interest in automatic facial age estimation. In this paper, different age variation methods with their prospects are reviewed. This paper highlights latest methodologies and feature extraction methods used by researchers to estimate age. Different types of classifiers used in this domain have also been discussed.

Mortality Estimation: Biased Results from Unbiased Ages

1984 ◽  
Vol 41 (12) ◽  
pp. 1843-1847 ◽  
Author(s):  
Jay Barlow
Keyword(s):  
Monte Carlo ◽  
Monte Carlo Simulations ◽  
Age Estimation ◽  
Age Determination ◽  
Mortality Rates ◽  
Age Classes ◽  
Estimation Errors ◽  
Mortality Estimation ◽  
Age Estimates

Estimates of mortality rates from age distributions are biased by imprecision in age estimation, even if age estimates are unbiased. I have derived a method for predicting the magnitude of this bias from information on the precision of age determination. Monte Carlo simulations show that bias can be accurately predicted. The commonly used Chapman–Robson mortality estimator is shown to be robust to imprecision in age determination if all age-classes are included. Errors are likely, however, if one or more age-classes are excluded or if other mortality estimators are used. Biases can be corrected if the distribution of age-estimation errors is known.

Death and Society in Twentieth Century America

1988 ◽  
Vol 18 (4) ◽  
pp. 379-395 ◽  
Author(s):  
Robert Fulton ◽  
Greg Owen
Keyword(s):  
Twentieth Century ◽  
Research Agenda ◽  
Atomic Bomb ◽  
Age Groups ◽  
Time Span ◽  
Nuclear Age ◽  
The Media ◽  
Immediate Experience ◽  
Twentieth Century America ◽  
Two Age Groups

American attitudes and responses toward death have changed markedly during the twentieth century. This transformation is illustrated through an examination of two age groups: those born prior to the advent of the atomic bomb, and those born into the nuclear age. Each cohort contended with very different patterns of environment and socio-historical experiences, and had differential life expectancies as well. Images of death have changed significantly over this time-span, partially because of the pervasive influence of television and the overall growth in the importance of media. Death's presence in the media is simultaneously everywhere and nowhere; it is at once illusively fantastical and frighteningly real. Today's youth face the threat of a sudden anonymous death that is counterpoised against a more immediate experience with death that often is either distorted or denied. It is within this context that America's youth express their fears and frustrations in music, drugs, violence, and vicarious death experiences. The research agenda should include investigation of such phenomena as the rising interest in spirituality and the increase in suicide among adolescents as possible symptoms of despair in an impersonal and threatened world.

Facial Age Estimation by Learning Label Distribution CNN

2021 ◽  
Author(s):  
Kuan-Hsien Liu ◽  
Chun-Te Chang ◽  
Tsung-Jung Liu
Keyword(s):  
Age Estimation ◽  
Label Distribution ◽  
Facial Age
Sign in / Sign up

Export Citation Format

Share Document