Ready Player Me’s Post

Just under 4 years ago, Ready player me was slowly making its entry. Our VR companies were solicited, then on a one-to-one basis, we had a quick call with the co-founders, and we were partners in what is becoming the royal roadmap for standardized avatars for the future interconnection of 3D worlds, still called the “Metavers”. It's clear that Kaspar Tiri and Timmu Tõke's vision is in line with a well-written and clearly laid-out cultural strategy. After reaching a significant user peak, they serenely raised funds, and transformed their MVP with feedback from their users, to offer this year multiple simple and evolved features, accelerated thanks to AI and their know-how. Every week, we unveiled new features, transforming trials into market conquest. We're moving from similar avatars to the possibility of creating our own assets in a minimum of time, taking the standard from the visible to the invisible, in other words, having our own graphic style while keeping the “mechanical” structure identical, the royal road to standardization. Bravo team! Translated with DeepL.com (free version)

Introducing SceneScript Today, Meta #RealityLabs Research is announcing #SceneScript, a novel method of generating scene layouts and representing scenes using language. Rather than using hard-coded rules to convert raw visual data into an approximation of a room’s architectural elements, SceneScript is trained to directly infer a room’s geometry using end-to-end machine learning. This results in a representation of physical scenes which is #compact, reducing memory requirements to only a few bytes; #complete, resulting in crisp geometry, similar to scalable vector graphics; and importantly, #interpretable, meaning that we can easily read and edit those representations. https://lnkd.in/gs-tVEdH

📃Scientific paper: Planner3D: LLM-enhanced graph prior meets 3D indoor scene explicit regularization Abstract: Compositional 3D scene synthesis has diverse applications across a spectrum of industries such as robotics, films, and video games, as it closely mirrors the complexity of real-world multi-object environments. Conventional works typically employ shape retrieval based frameworks which naturally suffer from limited shape diversity. Recent progresses have been made in object shape generation with generative models such as diffusion models, which increases the shape fidelity. However, these approaches separately treat 3D shape generation and layout generation. The synthesized scenes are usually hampered by layout collision, which suggests that the scene-level fidelity is still under-explored. In this paper, we aim at generating realistic and reasonable 3D indoor scenes from scene graph. To enrich the priors of the given scene graph inputs, large language model is utilized to aggregate the global-wise features with local node-wise and edge-wise features. With a unified graph encoder, graph features are extracted to guide joint layout-shape generation. Additional regularization is introduced to explicitly constrain the produced 3D layouts. Benchmarked on the SG-FRONT dataset, our method achieves better 3D scene synthesis, especially in terms of scene-level fidelity. The source code will be released after publication. ;Comment: 16 pages, 10 figures Continued on ES/IODE ➡️ https://etcse.fr/k9G ------- If you find this interesting, feel free to follow, comment and share. We need your help to enhance our visibility, so that our platform continues to serve you.

🤩 "Rather than using hard-coded rules to convert raw visual data into an approximation of a room’s architectural elements, SceneScript is trained to directly infer a room’s geometry using end-to-end machine learning. This results in a representation of physical scenes which is compact, reducing memory requirements to only a few bytes; complete, resulting in crisp geometry, similar to scalable vector graphics; and importantly, interpretable, meaning that we can easily read and edit those representations." #research #meta #metaquest #computervision #ai #artificialintelligence #3d #augmentedreality #mixedreality

Another interesting image-image #AI tool, this time for #C4D. I think this sort of tooling is much better for artists and designers because it actually saves them time doing something tedious instead of trying to replace them outright. They can do a bunch of concepting/moodboarding very quickly at a low-resolution right in Cinema4D, get sign off on a particular direction and then execute at the highest level with their favorite renderer, probably Redshift or Octane in this case.

2015 - I just started studying Corona render and gaining access to the alpha version, I began to make a concept render of the “City of the Future” (these were my first steps in visualization). I spent several days setting up the environment, lighting, materials, textures, and other effects. 2024 - I spent a few minutes, 1-2 AI tools, got ~100 visualization options (day, night, sunset, retained the original geometry, more futuristic shapes). It's just a new tool for #nextlevelrendering

Now you can generate AI images inside of Blender, and it will take into account the placement of your 3D models. Today I was messing around with Stable Diffusion, and discovered that there’s a Blender script that allows you to create AI generated images directly from the 3D software. Obviously the examples in my video are nowhere close to being usable at all. The shapes are twisted, the body proportions are bad and the plane in the background is just weird. But being able to give Stable Diffusion a set of 3D models and let it enhance your render with some extra detail is not something to overlook. Do you guys know any similar resources to this one?

A NeRF-rendered scene looks incredibly realistic, capturing fine details, textures, and lighting effects in a 3D environment. It revolutionizes computer graphics by generating photorealistic images from 2D inputs. #nerf #3drendering #computergraphics #photorealism #ai #machinelearning #neuralnetworks #deeplearning #3dmodeling #TechInnovation #artificialintelligence #graphics #rendering #techrevolution #neuralrendering #3dvisualization #visualcomputing #cgi #neuralradiancefields #3dgraphics

More physics by Daniel Escobar. Last hours to join our upcoming workshop, “PixelSpace: Intro to 3D Generative AI” which will focus on the automation of 3D scene generation from text and images. The workshop will led by creative designer and 3D generative AI researcher Daniel Escobar (@daniel.esco1) co-founder of the @diffusion_architecture, on August 31 – September 1, 2024. Register Now: Tap the🔗 link: https://lnkd.in/eGhXyrpr This workshop will explore the cutting-edge techniques of multiview diffusion models conditioned on camera paths for 3D scene generation. We will dive into the latest methods for representing 3D scenes and geometry and investigate how current research in 3D generative AI leverages pre-trained image and video models for a 3D generation. Participants will learn how to use text and image inputs to generate scenes, which will then be imported into Unreal Engine for post-production and to create a concept reel. 📑 Topic: PixelSpace: Intro to 3D Generative AI 📅 Date: August 31 – September 1, 2024 🕕 Time: 15:00 - 19:00 GMT ⚒️ Software: Blender, Unreal Engine, Luma 🧑🏼🎓 Total Seats: 50 Seats 🛒 15% Discount for Digital Members. #artificialintelligence #midjourney #parametricdesign #computationaldesign #architecturestudents #formgeneration #unrealengine #blender