Abstract. Most computational models of gender classification use global information (the full face image) giving equal weight to the whole face area irrespective of the importance of the internal features. Here we use a two-way representation of face images that includes both global and featural information. We use dimensionality reduction techniques and a support vector machine classifier and show that this method performs better than either global or feature based representations alone. 1

We apply the feature-selection technique of Fukunaga and Koontz, an extension of the Karhunen-Lo`eve transformation, to spoken letter recognition. Feedforward networks trained for letter-pair discrimination with the new features show up to 37% reduction in classifier error rate relative to networks trained with spectral coefficients. This performance increase is accompanyed by a 91% reduction in feature dimension. For three-letter discrimination, the new features perform comparably to spectral coefficients, with a 90% reduction in feature dimension. 1 Introduction It is generally recognized that compact feature sets ameliorate the computational burden for classifiers. Researchers have used algebraic and neural network implementations of principal component analysis to compress high-dimensional data for speech [1], acoustic emission [2], and vision [3, 4] problems. Beyond compression, statistical feature-selection techniques should provide features with good discriminatory powe...

This paper introduces an automatic real-time gender classification system. The system consists of mainly three modules, face detection, normalization and gender classification. The LUT-type weak classifier based Adaboost learning method is proposed for training both face detector and gender classifier, and a Simple Direct Appearance Model (SDAM) based method is developed to detect the facial landmark points for face normalization. This results in an integrated system with rather good performance. Experiment results on both pictures from World Wide Web and real-time video clips are reported to demonstrate its effectiveness and robustness.

Human facial images provide the demographic information, such as ethnicity and gender. Conversely, ethnicity and gender also play an important role in face-related applications. Image-based ethnicity identification problem is addressed in a machine learning framework. The Linear Discriminant Analysis (LDA) based scheme is presented for the two-class (Asian vs. non-Asian) ethnicity classification task. Multiscale analysis is applied to the input facial images. An ensemble framework, which integrates the LDA analysis for the input face images at di#erent scales, is proposed to further improve the classification performance. The product rule is used as the combination strategy in the ensemble. Experimental results based on a face database containing 263 subjects (2,630 face images, with equal balance between the two classes) are promising, indicating that LDA and the proposed ensemble framework have su#cient discriminative power for the ethnicity classification problem. The normalized ethnicity classification scores can be helpful in the facial identity recognition. Useful as a "soft" biometric, face matching scores can be updated based on the output of ethnicity classification module. In other words, ethnicity classifier does not have to be perfect to be useful in practice.