Meta 3D Gen AI System
04-07-2024
10:13 AM
1 min read
Overview:
Recently, Meta has introduced a new AI tool that can generate or retexture 3D objects in under a minute.
About Meta 3D Gen AI System:
- It is a combined AI system that can generate high-quality 3D assets, with both high-resolution textures and material maps, all within one minute.
- It uses two AI-powered sub-systems called Meta 3D AssetGen and Meta 3D TextureGen.
- The system can take text prompts from the user and produce a final output of AI generated 3D content such as characters, props or scenes.
- Additionally, the user can also provide the system with an already created 3D mesh, which is essentially the structure of a 3D model, and ask it to add texture to it.
- How does it works?
- This system uses Physically-Based Rendering (PBR) techniques to build up 3D content from scratch.
- The system first takes an input from the user which contains instructions for creating a 3D structure and its design elements including the texture.
- In the first step, the Meta 3D AssetGen system creates an initial “3D asset” based on the user prompt. This step produces a 3D mesh with texture and PBR material maps. The inference time for this step is approximately 30 seconds.
- In stage two, a higher-quality texture and PBR maps are generated for the 3D asset generated in stage one. This step utilises the Meta 3D TextureGen system that takes approximately 20 seconds, according to the company.
- Additionally, the Meta 3D TextureGen system can be used individually by the user as a text-to-texture generator. If the user provides the system with an already generated 3D mesh along with a text prompt detailing the texture they want on the 3D content, the system automatically initiates step two.
Q1: What is Generative Artificial Intelligence?
It refers to deep-learning models that can take raw data and “learn” to generate statistically probable outputs when prompted. It is powered by foundation models (large AI models) that can multi-task and perform out-of-the-box tasks, including summarization, Q&A, classification, and more.