Trace4D: Grounded Spatiotemporal Reasoning via 4D State Sequences
cs.CV
Submitted: 2026-11-02
Updated: 2026-11-02
Project page: https://cvlab-kaist.github.io/Trace4D
Terminology
Sources
- LLaVA-OneVision-1.5: Fully Open Framework for Democratized Multimodal Training
- Qwen3-VL Technical Report
- Qwen2.5-VL Technical Report
- 4DThinker: Thinking with 4D Imagery for Dynamic Spatial Understanding
- Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling
- 4DP-QA: Scalable QA for 4D Perception in Vision Language Models
- DynTrace: Tracking Dynamic Object Evidence for 4D Spatio-Temporal Reasoning in MLLMs
- LoRA: Low-Rank Adaptation of Large Language Models
- GPT-4o System Card
- ST-VLM: Kinematic Instruction Tuning for Spatio-Temporal Reasoning in Vision-Language Models
- GeoAlign: Geometric Feature Realignment for MLLM Spatial Reasoning
- Open-o3-Video: Grounded Video Reasoning with Explicit Spatio-Temporal Evidence
- GeoWeaver: Grounding Visual Tokens with Geometric Evidence before Scene Reasoning
- SpaceR: Reinforcing MLLMs in Video Spatial Reasoning
- Grounded SAM: Assembling Open-World Models for Diverse Visual Tasks
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- OpenAI GPT-5 System Card
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- Video-STR: Reinforcing MLLMs in Video Spatio-Temporal Reasoning with Relation Graph
- Cambrian-P: Pose-Grounded Video Understanding
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs