Details

• Led research and design of an AI-powered 3D workflow at Upland (Animoca Brands portfolio company).
• Implemented the workflow across teams, transforming the 3D production pipeline.
• Reduced production time by 72%.

Text-to-Image Prompt

The workflow begins with different text-to-image prompt solutions (Stable Diffusion 1.5, Midjourney, Lexica), empowering Upland's team to generate diverse AI-sketches.

Mesh Generation

The workflow proceeds by employing Image-to-Mesh AI solutions, including Kaedim, CSM, Layer, and Meshy, to produce a refined 3D mesh model.

Clean Up

In the last step of the workflow, the 3D model is refined, and the original input texture imagery is projected onto the mesh.

Render Output

The final output is ready for in-game production. 🤙