My primary contribution was to the manual portion of the process, specifically to keep the characters' faces consistent. AI helped with generation, but it was not accurate enough to maintain the characters' visual consistency between pictures; hence, manual correction and refinement were crucial. To guarantee visual coherence and quality, the editing and final polish were also completed by hand. Some of the pictures produced for the Zepto project are shown here.
The shot that you see below was created by me as part of Project Zepto, a team-based production during my time at AVGC. The project was a blend of AI-driven workflows and manual craftsmanship.
These images weren't included in the episode since the customer requested that the background be altered, but I included them because I preferred this background.
Here is a before and after photo of the input I used to produce the artwork.
Even though this was a team effort, I worked on the shots that are displayed here. Being a part of this cooperative effort was quite enjoyable for me. and I'm currently working on the 2D episode. around 80% of the second episode's pictures. The pictures from the second episode are shown below.