AI Research 技术调研

标签: iccv2025

此标签下有6条笔记。

  • 2026年7月16日

    HERMES: A Unified Self-Driving World Model for Simultaneous 3D Scene Understanding and Generation

    • driving-world-model
    • bev-representation
    • world-queries
    • internvl2
    • point-cloud-forecasting
    • scene-understanding
    • unified-understanding-generation
    • nuscenes
    • iccv2025
  • 2026年7月16日

    I2-World: Intra-Inter Tokenization for Efficient Dynamic 4D Scene Forecasting

    • world-model
    • autonomous-driving
    • 4d-occupancy
    • tokenization
    • residual-quantization
    • occupancy-forecasting
    • nuscenes
    • occ3d
    • iccv2025
  • 2026年7月16日

    TesserAct: Learning 4D Embodied World Models

    • world-model
    • 4d-scene-reconstruction
    • rgb-depth-normal
    • video-diffusion
    • cogvideox
    • action-conditioned-prediction
    • robotic-manipulation
    • inverse-dynamics
    • point-cloud
    • iccv2025
  • 2026年7月16日

    World4Drive: End-to-End Autonomous Driving via Intention-aware Physical Latent World Model

    • world-model
    • autonomous-driving
    • end-to-end-planning
    • latent-world-model
    • self-supervised
    • vision-foundation-model
    • multi-modal-trajectory
    • nuscenes
    • navsim
    • iccv2025
  • 2026年6月25日

    Lumina-Image 2.0: A Unified and Efficient Image Generative Framework

    • t2i
    • dit
    • flow-matching
    • unified-attention
    • gemma2
    • recaptioning
    • open-source
    • iccv2025
  • 2026年6月25日

    VACE: All-in-One Video Creation and Editing

    • video
    • editing
    • unified
    • dit
    • controllable
    • inpainting
    • reference-to-video
    • wan
    • ltx-video
    • iccv2025

  • GitHub