Human faces provide crucial information regarding gender, age, and ethnicity, in addition to identity. Several important fields for applications of gender classification have been identified, such as biometric authentication, surveillance and security systems, demographic information collection, marketing research, real time electronic marketing, criminology, augmented reality, and lately, new applications in social networks using face recognition. Gender classification based on facial images is currently one of the most challenging problems in image analysis research.
In image understanding, raw input data often has very high dimensionality and a limited number of samples. In this area, feature selection plays an important role in improving accuracy, efficiency and scalability of the object identification process. Since relevant features are often unknown a priori in the real world, irrelevant and redundant features may be introduced to represent the domain. However, using more features implies increasing computational cost in the feature extraction process, slowing down the classification process and also increasing the time needed for training and validation, which may lead to classification over-fitting.
As is the case in most image analysis problems, with a limited amount of sample data, irrelevant features may obscure the distributions of the small set of relevant features and confuse the classifier. It has been shown both theoretically and empirically that reducing the number of irrelevant or redundant features significantly increases the learning efficiency of the classifier.
We can see in the video the most relevant features on the female and male faces selected from the fusion of different features extraction methods: