A comprehensive study of the processes involved in developing and optimizing character designs for top-down games, from concept to completion

Abstract

This thesis investigates the integration of generative artificial intelligence into the production pipeline of a stylized character for a top-down game. The study addresses the following research question: How generative AI tools can be incorporated into character production workflows, and how this integration affects production efficiency, controllability, and visual quality. The research follows a practice-based approach and is structured as a case study involving the development of a stylized character asset for a top-down game context. The production pipeline combines generative AI tools with traditional game development software and includes concept generation, visual refinement, three-dimensional modelling, optimization, and animation preparation. The process is systematically documented and evaluated based on defined criteria. In addition, a questionnaire with 16 artists was conducted to assess the perceived quality of the final character. The results indicate that AI-assisted workflows can significantly accelerate early stages of production, particularly during concept development and initial asset creation. At the same time, limitations in controllability and stylistic consistency were observed, requiring iterative refinement and human intervention. The findings suggest that generative AI is most effective as a supporting tool in hybrid workflows, accelerating early stages while requiring human control over artistic direction and final quality.