# MoVISA: Multi-Token Reasoning for Video Object Segmentation

arXiv: https://arxiv.org/abs/2609.28956 · HTML: https://mlsanity.com/paper/2609.28956/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 24 of its 30 references that are in the map.
- First posted on arXiv 24 Sep 2026.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [Video understanding with MLLMs](https://mlsanity.com/mllm/video-understanding-with-mllms/)

## Builds on (most cited first, 24 of 24)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [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)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [Modeling Context in Referring Expressions](https://mlsanity.com/paper/1608.00272/) (2016)
- [Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models](https://mlsanity.com/paper/2306.05424/) (2023)
- [Otter: A Multi-Modal Model with In-Context Instruction Tuning](https://mlsanity.com/paper/2305.03726/) (2023)
- [LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models](https://mlsanity.com/paper/2311.17043/) (2023)
- [MAttNet: Modular Attention Network for Referring Expression Comprehension](https://mlsanity.com/paper/1801.08186/) (2018)
- [MeViS: A Large-scale Benchmark for Video Segmentation with Motion Expressions](https://mlsanity.com/paper/2308.08544/) (2023)
- [Universal Instance Perception as Object Discovery and Retrieval](https://mlsanity.com/paper/2303.06674/) (2023)
- [Language as Queries for Referring Video Object Segmentation](https://mlsanity.com/paper/2201.00487/) (2022)
- [Video Object Segmentation with Language Referring Expressions](https://mlsanity.com/paper/1803.08006/) (2018)
- [End-to-End Referring Video Object Segmentation with Multimodal Transformers](https://mlsanity.com/paper/2111.14821/) (2021)
- [VISA: Reasoning Video Object Segmentation via Large Language Models](https://mlsanity.com/paper/2407.11325/) (2024)
- [OnlineRefer: A Simple Online Baseline for Referring Video Object Segmentation](https://mlsanity.com/paper/2307.09356/) (2023)
- [Language-Bridged Spatial-Temporal Interaction for Referring Video Object Segmentation](https://mlsanity.com/paper/2206.03789/) (2022)
- [GLUS: Global-Local Reasoning Unified into A Single Large Language Model for Video Segmentation](https://mlsanity.com/paper/2504.07962/) (2025)
- [The Devil is in Temporal Token: High Quality Video Reasoning Segmentation](https://mlsanity.com/paper/2501.08549/) (2025)
- [TrackGPT -- A generative pre-trained transformer for cross-domain entity trajectory forecasting](https://mlsanity.com/paper/2402.00066/) (2024)

## Abstract (arXiv)

Recent advances in video object segmentation with Multimodal Large Language Model (MLLM) reasoning have demonstrated the effectiveness of using a single textual token, such as SEG, to predict segmentation masks across images and videos. However, we observe that this single-token strategy lacks the granularity required to precisely localize multiple objects across time in video segmentation tasks. To address this limitation, we develop Multi-Token Reasoning for Video Object Segmentation, or MoVISA. MoVISA uses multiple segmentation tokens, such as SEG0 and SEG1, to represent an object across different frames. This design enables more fine-grained alignment between language prompts and spatio-temporal mask predictions, improving both performance and interpretability. On the challenging MeViS, DAVIS17, ReVOS, and Ref-Youtube-VOS benchmarks, our model achieves a 13.2 percent J and F improvement on MeViS and an 8.4 percent J and F improvement on ReVOS. Code and models will be released.