Gender Classification of Face Images: The Role of Global and Feature-Based Information

Abstract

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.

The Role of Face Parts in Gender Recognition

“…For example, the relation between gender recognition rates from eyes and mouth is inverted in these two papers, eyes being more discriminant than mouth in [4], and mouth more accurate than eyes in [3]. This paper goes beyond these two previous related works, as regards the number of face parts and the diversity of the experimental design.…”

“…Even in cases in which exogenous and cultural features such us hair and makeup are removed, adults are able to correctly recognize gender in about 96% of cases. In recent years, a number of papers have paid attention to the problem of automatic gender recognition from face image inspection [1][2][3][4]. Possible applications can be imagined in access control, in the demographic description of a population, and so forth.…”

“…Two recent papers [3,4] have studied the importance of face parts in gender recognition. In [3], the differentiation capabilities of full face, jaw, lips/mouth, nose and eyes were evaluated.…”

“…Other closely related work is [4], where the roles of the full face, eyes and mouth in gender classification were compared. The subimages of internal features (eyes and mouth) were extracted and their dimensionality reduced using PCA, CCA and SOM.…”

“…Considering that results about the discriminant capabilities of face parts obtained in [3,4] depend on specific components (dataset, classifier, face part description), care should be taken when interpreting their conclusions. For example, the relation between gender recognition rates from eyes and mouth is inverted in these two papers, eyes being more discriminant than mouth in [4], and mouth more accurate than eyes in [3].…”