OpenAI releases Point-E, an AI that generates 3D models

December 21, 2022 By Raulf Hernes

(Image Credit Google)

OpenAI has released a machine learning system that can generate a 3D object from a text prompt. On a single Nvidia V100 GPU, Point-E can generate 3D models in one to two minutes. It creates point clouds, or discrete data points in space, rather than 3D objects. From a computational standpoint, point clouds are easier to create. To solve this issue, the Point-E team instructed a second AI system to transform the point clouds in Point-E to meshes. (Meshes are collections of vertices, edges, and faces that define an object in 3D modelling and design.) However, the researchers note in their paper that the model can occasionally miss specific parts of objects, tend to result in blocky or distorted shapes. Apart from the mesh-generating model, Point-E has two models: a text-to-image model and an image-to-3D model. The text-to-image model, which is similar to OpenAI's own DALL-E 2 and Stable Diffusion generative art systems, was trained on labelled images to understand the links between words and visual concepts. The image-to-3D model, while on the other hand, was fed a set of pictures paired with 3D objects so that it could learn to translate between the two effectively. [caption id="attachment_74329" align="aligncenter" width="1750"] 3D Models by AI

OpenAI[/caption] When given a text prompt, such as "a 3D printable gear, a single gear 3 inches in diameter and half inch thick," Point-text-to-image E's model creates a synthetic rendered object, which is then fed into the image-to-3D model, which creates a point cloud. According to the OpenAI researchers, after training the models on a dataset of "several million" 3D objects and associated metadata, Point-E could generate coloured point clouds that consistently matched text prompts. It isn't perfect; Point-image-to-3D E's model occasionally fails to recognize the image from the text-to-image model, eventually results in a shape that does not match the text prompt. Nonetheless, according to the OpenAI team, it is orders of magnitude faster than the previous state-of-the-art. The software could also be integrated into game and animation development workflows.OpenAI is the most recent company to enter the 3D object generator fray. Google DreamFusion, an enhanced version of Dream Fields, was released earlier this year. 3D models are commonly used in film and television, interior design, architecture, and a variety of scientific fields. They are used by architectural firms to demonstrate proposed buildings and landscapes. Engineers use them to create new devices, vehicles, and structures.

By Raulf Hernes

If you ask me raulf means ALL ABOUT TECH!!

OpenAI releases Point-E, an AI that generates 3D models

Walmart's GenAI Search Engine: Challenging Google's Dominance in Retail Search

Introducing the MSI Claw Handheld: Your Ultimate Gaming Companion

Uber Increases In-App Ads, Prompting Mixed Reactions from Customers

Apple's iOS 18: A Leap into the AI Era

Google's Regular Pixel 8 Won't Get Gemini Nano AI

MacBook Air M3 Makes Amends for M2's Storage Blunder

Samsung Unveils the Galaxy M15 5G

Elon Musk's xAI to Open-Source Chatbot Grok

Contra: Operation Galuga - A Modern Run-and-Gun Classic

Musk Confirms X's TV App Arrives This Week

Samsung Unveils the Galaxy M15 5G

Google Pixel Watch 3: Bigger, Better, and Packed with Potential

Google Pixel Phones in Line for SOS Feature via Satellite

Tech Giants Bet Big on Figure, the Humanoid Robot Company

Galaxy AI Takes Flight: Try Before You Buy on Any Phone!

Eufy Launches Wi-Fi-Free, 360° Security Camera with Built-in Solar Power