This research paper presents a novel machine learning based approach for generating personalized arms for virtual reality use cases. The approach is fully automatic and not bound to expensive or specialized hardware. To overcome the big amount of data necessary to train machine learning models a synthetic data generation scheme is employed. It is then shown how an image to image machine learning model can be trained via this training data to extract personalized arm textures from only two photographs. Finally the resulting virtual arms are analyzed by conducting an experiment to measure the embodiment of test subjects using these personalized arms in a virtual reality environment. This work represents a notable advancement in the integration of machine learning with virtual reality. It offers a promising step towards more user-centric and immersive VR experiences without necessitating high-end hardware. The findings serve as a foundation for future research aimed at refining and expanding the applicability of personalized virtual environments.

Top