ABOUT

This set of animations completed for TechOffice is focused on their developments with 2D image generation and new ways of creating training models for Artificial Intelligence. SORDI GAN (Generative Adversarial Network) uses a new approach to producing synthetic data and building a new level of contextual awareness in future hardware on edge applications. While previous AI models were trained using synthetic data extracted from 3D models through a process known as rendering, this new method uses two QC AI models that compete to generate and refine images depicting factory scenes. This technology's applications include visualizing language, novel HMI interfaces, visual instruction, and further integrations of AI visualisation.

Building upon SORDI GAN, the first project exploring several AI models functioning together through a new LAM interface has been dubbed Genesis. The animation communicates a novel workflow that allows management and staff to create their own datasets, modify images and control parameters to a level of detail that was otherwise incredibly difficult before GAN. It re-used motion design work from the GAN project, and presented a vision for a software package utilising the emerging technology.