Generating Generations - Automatic generation of 3D characters by simulation of genetic inheritance

Sammanfattning: This thesis proposes a model that can simulate multiple generations of genetic inheritance as a way of creating varied 3D content automatically. The proposed model should be less complex and less resource demanding than existing models while producing considerably better results than a model generating content by simple randomization. To achieve variation, the design of the base 3D model for character customization was based on different animation techniques, modifiers were based on anthropometry elements from advanced genetic simulations and known cognitive factors important for recognition. By use of ideas from other successful content generation and other implementations of 3D genetics, four genetic inheritance models where designed, implemented and evaluated. The results show that the proposed model produces very good and advanced looking results of genetic kinship, and can be used for generation of large numbers of content and all while using small amounts of resources.