Rozpoznání pohlaví za pomocí obrazů

Abstract

The point of this bachelor thesis is to present process of gender recognition based on image analysis, describe basic terms in the field of gender and face recognition, describe commonly used methods or ways of solutions within this field and to specify use of software library OpenCV which is focused on image analysis. At the same time it tries to specify this process and describes certain theoretical knowledge significant for this field and practically usable. Thesis specifies way of gender recognition in images by means of the three conventionally used methods: Eigenfaces, Fisherfaces and Local Binary Patterns Histograms. The issues of these methods are defined by theoretical part, personal proposal, implementation and functional testing. Created and tested solutions are function basis for activity in the area of gender recognition. Results of this thesis may be a quality basis for further testing or further developement in this field.