| Sa2VA: Marrying SAM2 with LLaVA for Dense Grounded Understanding of Images and Videos | ✓ Link | 75.1 | | | Sa2VA-26B (extra data) | 2025-01-07 |
| VideoSEG-O3: A Multi-turn Reinforcement Learning Framework for Reasoning Video Object Segmentation | ✓ Link | 74.1 | | | VideoSEG-O3-4B | 2026-06-05 |
| MPG-SAM 2: Adapting SAM 2 with Mask Priors and Global Context for Referring Video Object Segmentation | ✓ Link | 73.9 | 71.7 | 76.1 | MPG-SAM 2 | 2025-01-23 |
| Find First, Track Next: Decoupling Identification and Propagation in Referring Video Object Segmentation | ✓ Link | 73.7 | | | FindTrack++ (N=10) (extra data) | 2025-03-05 |
| RCoT-Seg: Reinforced Chain-of-Thought for Video Reasoning and Segmentation | ✓ Link | 72.5 | | | RCoT-Seg | 2026-05-08 |
| MomentSeg: Moment-Centric Sampling for Enhanced Referring Video Object Segmentation | | 72.3 | | | MomentSeg-7B | 2026-09-08 |
| SPOT: Spatiotemporal Prompt Optimization for Motion-Stabilized MLLM-Guided Video Segmentation | | 71.8 | | | SPOT-13B | 2026-06-03 |
| VideoLoom: A Video Large Language Model for Joint Spatial-Temporal Understanding | ✓ Link | 71.3 | | | VideoLoom-8B (extra data) | 2026-01-12 |
| The Devil is in Temporal Token: High Quality Video Reasoning Segmentation | ✓ Link | 71 | 69 | 73.1 | VRS-HQ (Chat-UniVi-13B) | 2025-01-15 |
| General Object Foundation Model for Images and Videos at Scale | ✓ Link | 70.6 | 68.2 | 72.9 | GLEE-Pro | 2023-12-14 |
| InterRVOS: Interaction-aware Referring Video Object Segmentation | ✓ Link | 70.5 | | | ReVIOSa-4B (standard RVOS, AML-only) (extra data) | 2025-06-03 |
| Towards Streaming Referring Video Segmentation via Large Language Model | ✓ Link | 70.5 | | | StreamingRVOS-4B | 2026-06-03 |
| VideoSEG-O3: A Multi-turn Reinforcement Learning Framework for Reasoning Video Object Segmentation | ✓ Link | 70.5 | | | VideoSEG-O3-2B | 2026-06-05 |
| Boosting Text-Driven Video Segmentation via Geometry-Aware Distillation | ✓ Link | 70.5 | | | GeoLaV | 2026-06-23 |
| Universal Instance Perception as Object Discovery and Retrieval | ✓ Link | 70.1 | 67.6 | 72.7 | UNINEXT-H | 2023-03-12 |
| Deforming Videos to Masks: Flow Matching for Referring Video Segmentation | ✓ Link | 69.6 | | | FlowRVS (Wan 2.1 T2V 1.3B DiT, RefCOCO pretrained) (extra data) | 2025-10-07 |
| ReferDINO: Referring Video Object Segmentation with Visual Grounding Foundations | | 69.3 | 67.0 | 71.5 | ReferDINO (Swin-B) | 2025-01-24 |
| SAMWISE: Infusing Wisdom in SAM2 for Text-Driven Video Segmentation | ✓ Link | 69.2 | | | SAMWISE (Hiera-B + RoBERTa) (extra data) | 2024-11-26 |
| HyperSeg: Hybrid Segmentation Assistant with Fine-grained Visual Perceiver | ✓ Link | 68.5 | | | HyperSeg-3B | 2025-06-11 |
| Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation | ✓ Link | 68.4 | 66.4 | 70.4 | MUTR | 2023-05-25 |
| Harnessing Vision-Language Pretrained Models with Temporal-Aware Adaptation for Referring Video Object Segmentation | | 67.6 | 65.3 | 69.8 | VLP (VLMo-L) | 2024-05-17 |
| InstructSeg: Unifying Instructed Visual Segmentation with Multi-modal Large Language Models | ✓ Link | 67.5 | | | InstructSeg (Mipha-3B) (extra data) | 2024-12-18 |
| Segment Every Reference Object in Spatial and Temporal Spaces | | 67.4 | 65.5 | 69.2 | UniRef-L (Swin-L) | 2023-01-01 |
| SOC: Semantic-Assisted Object Cluster for Referring Video Object Segmentation | ✓ Link | 67.3±0.5 | 65.3 | 69.3 | SOC (Joint training, Video-Swin-B) | 2023-05-26 |
| GLUS: Global-Local Reasoning Unified into A Single Large Language Model for Video Segmentation | ✓ Link | 67.3 | | | GLUS-A (Additional-SFT) (extra data) | 2025-04-10 |
| Temporally Consistent Referring Video Object Segmentation with Hybrid Memory | ✓ Link | 67.1 | 65.3 | 68.9 | HTR (Pre-training) | 2024-03-28 |
| Decoupling Static and Hierarchical Motion Perception for Referring Video Segmentation | ✓ Link | 67.1 | 65 | 69.1 | DsHmp (Video-Swin-Base) | 2024-04-04 |
| UniRef++: Segment Every Reference Object in Spatial and Temporal Spaces | ✓ Link | 66.9 | 64.8 | 69.0 | UniRef++-L | 2023-12-25 |
| VideoGLaMM: A Large Multimodal Model for Pixel-Level Visual Grounding in Videos | ✓ Link | 66.8 | | | VideoGLaMM (Phi-3 Mini 3.8B, RVOS fine-tuned) (extra data) | 2024-11-07 |
| Exploring Pre-trained Text-to-Video Diffusion Models for Referring Video Object Segmentation | ✓ Link | 66.5 | | | VD-IT (ModelScopeT2V, RefCOCO pre-trained) (extra data) | 2024-03-18 |
| ViLLa: Video Reasoning Segmentation with Large Language Model | ✓ Link | 66.5 | 64.6 | 68.6 | ViLLa | 2024-07-18 |
| Tracking Anything with Decoupled Video Segmentation | ✓ Link | 66.0 | | | DEVA (ReferFormer) | 2023-09-07 |
| Spectrum-guided Multi-granularity Referring Video Object Segmentation | ✓ Link | 65.7 | 63.9 | 67.4 | SgMg (Pre-training) | 2023-07-25 |
| GroPrompt: Efficient Grounded Prompting and Adaptation for Referring Video Object Segmentation | | 65.5 | 64.1 | 66.9 | GroPrompt | 2024-06-18 |
| STAC: Selective Spatiotemporal Aggregation and Compression for Video Reasoning Segmentation | ✓ Link | 65.2 | | | STAC | 2026-07-03 |
| Expression Prompt Collaboration Transformer for Universal Referring Video Object Segmentation | | 65 | 62.9 | 67.2 | EPCFormer (ViT-H) | 2023-08-08 |
| Universal Segmentation at Arbitrary Granularity with Language Instruction | ✓ Link | 64.9 | 62.8 | 67.0 | UniLSeg-100 | 2023-12-04 |
| Semantic and Sequential Alignment for Referring Video Object Segmentation | | 64.3 | | | SSA (extra data) | 2025-06-11 |
| Decoupled Motion Expression Video Segmentation | | 64.3 | | | DMVS | 2025-06-11 |
| LoSh: Long-Short Text Joint Prediction Network for Referring Video Object Segmentation | ✓ Link | 64.2 | 62.5 | 66.0 | LoSh-R | 2023-06-14 |
| VLT: Vision-Language Transformer and Query Generation for Referring Segmentation | ✓ Link | 63.8 | 61.9 | 65.6 | VLT | 2022-10-28 |
| One Token to Seg Them All: Language Instructed Reasoning Segmentation in Videos | ✓ Link | 63.7 | | | VideoLISA-3.8B (Post-optimization) (extra data) | 2024-09-29 |
| OnlineRefer: A Simple Online Baseline for Referring Video Object Segmentation | ✓ Link | 63.5 | 61.6 | 65.5 | OnlineRefer (Swin-L, online) | 2023-07-18 |
| VISA: Reasoning Video Object Segmentation via Large Language Models | ✓ Link | 63.0 | | | VISA Chat-UniVi-13B (extra data) | 2024-07-16 |
| Towards Robust Referring Video Object Segmentation with Cyclic Relational Consensus | ✓ Link | 61.3 | 59.6 | 63.1 | R2VOS (Video-Swin-T) | 2022-07-04 |
| SOC: Semantic-Assisted Object Cluster for Referring Video Object Segmentation | ✓ Link | 59.2 | 57.8 | 60.5 | SOC (Video-Swin-T) | 2023-05-26 |
| UniVS: Unified and Universal Video Segmentation with Prompts as Queries | ✓ Link | 58.0 | 56.8 | 59.5 | UniVS(Swin-L) | 2024-02-28 |
| Language as Queries for Referring Video Object Segmentation | ✓ Link | 57.3 | 56.1 | 58.4 | ReferFormer (ResNet-101) | 2022-01-03 |
| Decomposed Attention Fusion in MLLMs for Training-Free Video Reasoning Segmentation | ✓ Link | 57.0 | | | DecAF (training-free, Qwen2.5-VL-7B + SAM2) (extra data) | 2025-10-22 |
| Multi-Attention Network for Compressed Video Referring Object Segmentation | ✓ Link | 55.63 | 54.75 | 56.51 | MANET | 2022-07-26 |
| Language as Queries for Referring Video Object Segmentation | ✓ Link | 55.6 | 54.8 | 56.6 | ReferFormer (ResNet-50) | 2022-01-03 |
| End-to-End Referring Video Object Segmentation with Multimodal Transformers | ✓ Link | 55.32 | 54.00 | 56.64 | MTTR (w=12) | 2021-11-29 |
| Local-Global Context Aware Transformer for Language-Guided Video Segmentation | ✓ Link | 50 | 48.8 | 51.1 | Locater | 2022-03-18 |
| Multi-Level Representation Learning With Semantic Alignment for Referring Video Object Segmentation | | 49.70 | 50.96 | 48.43 | MLRLSA | 2022-01-01 |
| Deeply Interleaved Two-Stream Encoder for Referring Video Segmentation | | 49.56 | 48.44 | 50.67 | VLIDE | 2022-03-30 |
| URVOS: Unified Referring Video Object Segmentation Network with a Large-Scale Benchmark | ✓ Link | 48.9 | 47.0 | 50.8 | URVOS | |
| InternVideo2.5: Empowering Video MLLMs with Long and Rich Context Modeling | ✓ Link | 34.2 | | | InternVideo2.5 | 2025-01-21 |