Generative Visuals R&D
Prototyping and research on open-sourced generative visuals
Through a series of research and explorations into AI generative visual design, I gained a deeper understanding of the workflows required to achieve consistent and precisely controlled text-to-image and text-to-video outputs. This self-initiated generative project aimed to adopt AI as a creative medium to enhance and streamline my creative workflows for future projects. Below are the key research and development outcomes from this exploration:
Setting up both local and cloud-based computing nodes for generative AI to enable accessible and efficient workflows.
Training a LoRA model with set of my own photos to generate personalized images using Stable Diffusion 1.5.
Leveraging ControlNet for precise control and addressing video de-flickering to ensure cohesive outputs.
Developing generative AR video effects, generative QR codes, generative image-to-image pictures, generative self-portraits, generative voices
