This weekend in Generative Media

Disney Imagineering robots; China's training data rules; Google's genAI approach

• You can keep your Musk Optimus terror bot, I want a Disney Imagineering robot (TechRadar).

• China proposes blacklist of training data for generative AI models (ChannelNewsAsia)

• Our responsible approach to building guardrails for generative AI (Google blog)

• MotionDirector: Motion Customization of Text-to-Video Diffusion Models (project page)

• 4D Gaussian Splatting for Real-Time Dynamic Scene Rendering (project page)

• Idea2Img: Iterative Self-Refinement with GPT-4V(ision) for Automatic Image Design and Generation (project page)

• Instant3D: Fast Text-to-3D with Sparse-View Generation and Large Reconstruction Model (project page). High quality, diverse and Janus-free Text-to-3D in 20 SECONDS.

• LRM: Large Reconstruction Model for Single Image to 3D (project page). Create high-fidelity 3D object mesh from a single image in 5 SECONDS.

• HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion (project page)

• FLATTEN: optical FLow-guided ATTENtion for consistent text-to-video editing (project page)

• SyncDiffusion: Coherent Montage via Synchronized Joint Diffusions (NeurIPS 2023, project page)

• 3D animations for all, powered by AI (Anything World)

• Realize your vision with AI like never before (vispunk.com)

• Your masterpiece would look great on the big screen. Make it happen (Storia)

• Cuebric is putting the full-power of generative AI into the hands of the world's most visionary filmmakers and content creators. Walter Schulz shared with us below how he could re-create the opening scene of Prometheus in Cuebric. (LinkedIn)

• Another example of drawing to video (X)

• My ‘Glitch’ film is being screened at Sony Studios in LA next month. The promoters asked for a workflow of my process. I thought I’d also share it here. (X)

• Create a promotional image or title screen for a popular mobile game. (X)

• GPT-4V "Describe this image" 🔃 Dall-E 3 "Generate this image" Recursive loop (X)

• Just like ChatGPT can rewrite any text in any different style and length... ChatGPT Vision+Dall-E 3 can recreate an image in any other style. (X)

• 3D Gaussian Splats use case? We changed the opening scene of 'The Shining' after 43 years😮 (X)