2008
DOI: 10.1016/j.patrec.2008.03.016
|View full text |Cite
|
Sign up to set email alerts
|

An experimental comparison of gender classification methods

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
130
0
11

Year Published

2010
2010
2023
2023

Publication Types

Select...
4
2
1

Relationship

0
7

Authors

Journals

citations
Cited by 173 publications
(142 citation statements)
references
References 17 publications
0
130
0
11
Order By: Relevance
“…We used fa-and fb-subsets from the FERET database by removing duplicate images of the same subject; at the end there are 450 facial images for both females and males. For the MORPH database, which is not used in [1], we collect 8,033 female and 47,810 male images; there are totally 418 subjects of age ranging from 18-69 and of races of Caucasian, AfricanAmerican and Asian. To the best of our knowledge, we are the first tackling the gender identification problem over such variations with wide range of age and races.…”
Section: Resultsmentioning
confidence: 99%
See 1 more Smart Citation
“…We used fa-and fb-subsets from the FERET database by removing duplicate images of the same subject; at the end there are 450 facial images for both females and males. For the MORPH database, which is not used in [1], we collect 8,033 female and 47,810 male images; there are totally 418 subjects of age ranging from 18-69 and of races of Caucasian, AfricanAmerican and Asian. To the best of our knowledge, we are the first tackling the gender identification problem over such variations with wide range of age and races.…”
Section: Resultsmentioning
confidence: 99%
“…Previous studies on gender identification have relied on careful alignment of a facial image into a standard template. The studies in [1], [2] conduct experiments on many combination of the state-of-the-art face alignment and pattern classification methods for gender identification. The authors found that the classification rates increase with the accuracy of face alignment because such alignment could reduce the variability during building the model in the training phase.…”
Section: Introductionmentioning
confidence: 99%
“…This kind of approach has been shown to achieve high performance rates [4,1], but a strong preprocessing stage (normalization, alignment, scaling and/or histogram equalization) is needed. Based on the state of the art, we normalize and align the images using the eye positions, and rescale them to a resolution of 24 × 24 pixels.…”
Section: Classification Methods Using Normalized Imagesmentioning
confidence: 99%
“…Automatic gender classification of face images is an area of growing interest with multiple applications such as demographic data collection or facial expression recognition (for a recent review see [1]). In most practical applications, the properties of the images that will be used in the exploitation phase (illumination, pose, scale, orientation, occlusions, etc.)…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation