Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Evaluation of Gender Classification Methods with Automatically Detected and Aligned Faces Erno Ma ¨kinen and Roope Raisamo Abstract —We present a systematic study on gender classification with automatically detected and aligned faces. We experimented with 120 combinations of automatic face detection, face alignment, and gender classification. One of the findings was that the automatic face alignment methods did not increase the gender classification rates. However, manual alignment increased classification rates a little, which suggests that automatic alignment would be useful when the alignment methods are further improved. We also found that the gender classification methods performed almost equally well with different input image sizes. In any case, the best classification rate was achieved with a support vector machine. A neural network and Adaboost achieved almost as good classification rates as the support vector machine and could be used in applications where classification speed is considered more important than the maximum classification accuracy. Index Terms —Classifier design and evaluation, computer vision, face and gesture classification, face detection, interactive systems, machine learning, vision I/O. Ç 1I NTRODUCTION IN principle, it is fairly easy to combine face detection and gender classification methods. The face is detected with the detector and then input into the gender classifier that determines the gender. However, this process is more complex than it appears and includes many aspects for consideration. The most important factors are usually the detection and classification accuracies. The other important factors are detection and classification speeds. The issues to be considered are the selection of the detector and classifier, the features that are input of the detected face to the gender classifier, which (if any) normalization is used before gender classification, and if there is some processing that can be made common for both detection and classification. We examined the connection between face detection and gender classification experimentally. In the next section, we describe the related work. Then, the combination of the detector and classifier is considered from the technical perspective. Next, the experiments and their results are described and the results of the experiments are discussed. Finally, we present some concluding remarks. 2R ELATED WORK There exist only a few studies where gender classification has been combined with automatic face detection. These studies have proposed novel methods for gender classification. We contribute by evaluating combinations of gender classifiers and alignment methods where faces have been automatically detected. To the best of our knowledge, Moghaddam and Yang [1] developed the first automatic system for combined face detection and gender classification. They used maximum-likelihood estima- tion for face detection and for facial feature detection. For gender classification,theyusedseveraldifferentclassifiers.Theexperiments
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 10/23/2010 for the course COMMINUCAT 123 taught by Professor Ali during the Spring '10 term at Masaryk University.

Page1 / 7


This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online