ReVEAL: Eliciting Role-Aware Latent Visual Evidence from MLLMs for Reasoning Video Object Segmentation
cs.CV
Submitted: 2026-10-30
Updated: 2026-10-30
Terminology
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- Dreaming in Flow: Bidirectional Self-Refinement for Unified Multimodal Models
- Vorch-IR: Long-Form Unified Multimodal Identity Replacement Video Generation
- WorldMesh: Generating Navigable Multi-Room 3D Scenes via Mesh-Conditioned Image Diffusion
- OctWorld: Long-Range World-Consistent Video Generation with Octree-Based 3D Mapping
- Vorch-Streamer: Extending Human Audio-Visual Generation to Real-Time Long-Form Streaming