Stable Virtual Camera achieves state-of-the-art results in novel view synthesis (NVS) benchmarks, outperforming models like ViewCrafter and CAT3D. It excels in both large-viewpoint NVS, which emphasizes generation capacity, and small-viewpoint NVS, which prioritizes temporal
MULTIMODAL AI
-
AI Video Generation: Dynamic 3D Camera Control from Single Images
By
–
Key Capabilities: Dynamic Camera Control: Supports custom & preset paths (360°, Lemniscate, Dolly Zoom, Spiral, Pan, Move & more). Flexible Inputs: Generates 3D videos from just one input image or up to 32. Multiple Aspect Ratios: Produces square, portrait, landscape &
-
Stable Virtual Camera: 2D to 3D Video Transformation Model
By
–
Introducing Stable Virtual Camera: This multi-view diffusion model transforms 2D images into immersive 3D videos with realistic depth and perspective—without complex reconstruction or scene-specific optimization. pic.twitter.com/pHPkYhaKH3
— Stability AI (@StabilityAI) 18 mars 2025Introducing Stable Virtual Camera: This multi-view diffusion model transforms 2D images into immersive 3D videos with realistic depth and perspective—without complex reconstruction or scene-specific optimization.
-
OmniParser: Vision-Based GUI Agent by Microsoft
By
–
📷OmniParser: Vision Based GUI Agent
— Satya Mallick (@LearnOpenCV) 18 mars 2025
Microsoft's OmniParser represents a breakthrough in how AI systems interact with graphical interfaces. By using a pure vision-based approach to extract structured elements from UI screenshots, OmniParser helps large multimodal models like… pic.twitter.com/3qBzaxWV3IOmniParser: Vision Based GUI Agent Microsoft's OmniParser represents a breakthrough in how AI systems interact with graphical interfaces. By using a pure vision-based approach to extract structured elements from UI screenshots, OmniParser helps large multimodal models like
-
FlowMo: Image Tokenizer with Diffusion Autoencoder
By
–
Excited to share FlowMo – An image tokenizer with a diffusion autoencoder!🤩🎉 https://t.co/SS038d30tZ
— Fei-Fei Li (@drfeifei) 18 mars 2025Excited to share FlowMo – An image tokenizer with a diffusion autoencoder!
-
Token Injection Enables Layout-Guided Generative AI Image Composition
By
–
This new method enables: (a) layout-guided generation by translating token-injected regions, (b) multi-subject composition through sequential token injection, and (c) inpainting and outpainting via specifying masks and increased replacement.
-
Token Replacement with Patch Perturbations for Identity-Preserving Generation
By
–
During token replacement, variations are injected via patch perturbations. This timestep-adaptive strategy balances identity preservation and generative flexibility.
-
Personalize Anything: Mask-Guided Token Replacement for AI Models
By
–
Personalize Anything anchors subject identity in early denoising through mask-guided token replacement with preserved positional encoding, and transitions to multi-modal attention for semantic fusion with text in later steps.
-
Personalize Anything Free with Diffusion Transformer Technology
By
–
Tsinghua, Beihang, and RUC present: Personalize Anything for Free with Diffusion Transformer, enabling personalized image generation that produces user-specified concepts while allowing flexible editing.
-
Google Gemini May Roll Out Canvas Feature for Document and Code Interaction
By
–
BREAKING : Google may have started a rollout for its Canvas feature on Gemini to some users. So far, I've only seen one report from Japan. The Canvas feature will allow users to work with documents and code straight inside Gemini's response.