# One Trajectory, One Token: Grounded Video Tokenization via Panoptic Sub-object Trajectory

arXiv: https://arxiv.org/abs/2505.23617 · HTML: https://mlsanity.com/paper/2505.23617/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 60 of its 78 references that are in the map.
- First posted on arXiv 29 May 2025.

## Builds on (most cited first, 30 of 60)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [The Kinetics Human Action Video Dataset](https://mlsanity.com/paper/1705.06950/) (2017)
- [ViViT: A Video Vision Transformer](https://mlsanity.com/paper/2103.15691/) (2021)
- [Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis](https://mlsanity.com/paper/2405.21075/) (2024)
- [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://mlsanity.com/paper/2203.12602/) (2022)
- [Perceiver: General Perception with Iterative Attention](https://mlsanity.com/paper/2103.03206/) (2021)
- [The "something something" video database for learning and evaluating visual common sense](https://mlsanity.com/paper/1706.04261/) (2017)
- [Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models](https://mlsanity.com/paper/2306.05424/) (2023)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [Dense-Captioning Events in Videos](https://mlsanity.com/paper/1705.00754/) (2017)
- [NExT-QA:Next Phase of Question-Answering to Explaining Temporal Actions](https://mlsanity.com/paper/2105.08276/) (2021)
- [CLIP4Clip: An Empirical Study of CLIP for End to End Video Clip Retrieval](https://mlsanity.com/paper/2104.08860/) (2021)
- [DataComp: In search of the next generation of multimodal datasets](https://mlsanity.com/paper/2304.14108/) (2023)
- [An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models](https://mlsanity.com/paper/2403.06764/) (2024)
- [ActivityNet-QA: A Dataset for Understanding Complex Web Videos via Question Answering](https://mlsanity.com/paper/1906.02467/) (2019)
- [LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models](https://mlsanity.com/paper/2311.17043/) (2023)
- [MovieChat: From Dense Token to Sparse Memory for Long Video Understanding](https://mlsanity.com/paper/2307.16449/) (2023)
- [Hollywood in Homes: Crowdsourcing Data Collection for Activity Understanding](https://mlsanity.com/paper/1604.01753/) (2016)
- [Less is More: ClipBERT for Video-and-Language Learning via Sparse Sampling](https://mlsanity.com/paper/2102.06183/) (2021)
- [AVA: A Video Dataset of Spatio-temporally Localized Atomic Visual Actions](https://mlsanity.com/paper/1705.08421/) (2017)
- [Masked Autoencoders As Spatiotemporal Learners](https://mlsanity.com/paper/2205.09113/) (2022)
- [VATEX: A Large-Scale, High-Quality Multilingual Dataset for Video-and-Language Research](https://mlsanity.com/paper/1904.03493/) (2019)
- [Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers](https://mlsanity.com/paper/2402.19479/) (2024)

## Built on by (most cited first, 7 of 7)

- [SoccerMaster: A Vision Foundation Model for Soccer Understanding](https://mlsanity.com/paper/2512.11016/) (2025)
- [TrajTok: Learning Trajectory Tokens enables better Video Understanding](https://mlsanity.com/paper/2602.22779/) (2026)
- [Synthetic Visual Genome 2: Extracting Large-scale Spatio-Temporal Scene Graphs from Videos](https://mlsanity.com/paper/2602.23543/) (2026)
- [4D Primitive-Mach'e: Glueing Primitives for Persistent 4D Scene Reconstruction](https://mlsanity.com/paper/2512.16564/) (2025)
- [TrajPred: Trajectory-Conditioned Joint Embedding Prediction for Surgical Instrument-Tissue Interaction Recognition in Vision-Language Models](https://mlsanity.com/paper/2603.06999/) (2026)
- [Persistent Object Narratives for Token-Efficient Video Language Models](https://mlsanity.com/paper/2608.04866/) (2026)
- [Vision-Language Grounding as Bidirectional Concept Correspondence](https://mlsanity.com/paper/2608.07886/) (2026)

## Abstract (arXiv)

Effective video tokenization is critical for scaling transformer models for long videos. Current approaches tokenize videos using space-time patches, leading to excessive tokens and computational inefficiencies. The best token reduction strategies degrade performance and barely reduce the number of tokens when the camera moves. We introduce grounded video tokenization, a paradigm that organizes tokens based on panoptic sub-object trajectories rather than fixed patches. Our method aligns with fundamental perceptual principles, ensuring that tokenization reflects scene complexity rather than video duration. We propose TrajViT, a video encoder that extracts object trajectories and converts them into semantically meaningful tokens, significantly reducing redundancy while maintaining temporal coherence. Trained with contrastive learning, TrajViT significantly outperforms space-time ViT (ViT3D) across multiple video understanding benchmarks, e.g., TrajViT outperforms ViT3D by a large margin of 6% top-5 recall in average at video-text retrieval task with 10x token deduction. We also show TrajViT as a stronger model than ViT3D for being the video encoder for modern VideoLLM, obtaining an average of 5.2% performance improvement across 6 VideoQA benchmarks while having 4x faster training time and 18x less inference FLOPs. TrajViT is the first efficient encoder to consistently outperform ViT3D across diverse video analysis tasks, making it a robust and scalable solution.