Spatio-temporal localization is vital for precise interactions across diverse domains, from biological research to autonomous navigation and interactive interfaces. Current video-based approaches, while proficient in tracking, lack the sophisticated reasoning capabilities of large language models, limiting their contextual understanding and generalization. We introduce VideoMolmo, a large multimodal model tailored for fine-grained spatio-temporal pointing conditioned on textual descriptions. Building upon the Molmo architecture, VideoMolmo incorporates a temporal module utilizing an attention mechanism to condition each frame on preceding frames, ensuring temporal consistency. Additionally, our novel temporal mask fusion pipeline employs SAM2 for bidirectional point propagation, significantly enhancing coherence across video sequences. This two-step decomposition, i.e., first using the LLM to generate precise pointing coordinates, then relying on a sequential mask-fusion module to produce coherent segmentation, not only simplifies the task for the language model but also enhances interpretability. Due to the lack of suitable datasets, we curate a comprehensive dataset comprising 72k video-caption pairs annotated with 100k object points. To evaluate the generalization of VideoMolmo, we introduce VPoS-Bench, a challenging out-of-distribution benchmark spanning five real-world scenarios: Cell Tracking, Egocentric Vision, Autonomous Driving, Video-GUI Interaction, and Robotics. We also evaluate our model on Referring Video Object Segmentation (Refer-VOS) and Reasoning VOS tasks. In comparison to existing models, VideoMolmo substantially improves spatio-temporal pointing accuracy and reasoning capability. Our code and models are publicly available at https://github.com/mbzuai-oryx/VideoMolmo.
Long-horizon video-audio reasoning and fine-grained pixel understanding impose conflicting requirements on omnimodal models: dense temporal coverage demands many low-resolution frames, whereas precise grounding calls for high-resolution inputs. We tackle this trade-off with a two-system architecture: a Global Reasoning System selects informative keyframes and rewrites the task at low spatial cost, while a Detail Understanding System performs pixel-level grounding on the selected high-resolution snippets. Because ``optimal'' keyframe selection and reformulation are ambiguous and hard to supervise, we formulate them as a reinforcement learning (RL) problem and present Omni-R1, an end-to-end RL framework built on Group Relative Policy Optimization. Omni-R1 trains the Global Reasoning System through hierarchical rewards obtained via online collaboration with the Detail Understanding System, requiring only one epoch of RL on small task splits. Experiments on two challenging benchmarks, namely Referring Audio-Visual Segmentation (RefAVS) and Reasoning Video Object Segmentation (REVOS), show that Omni-R1 not only surpasses strong supervised baselines but also outperforms specialized state-of-the-art models, while substantially improving out-of-domain generalization and mitigating multimodal hallucination. Our results demonstrate the first successful application of RL to large-scale omnimodal reasoning and highlight a scalable path toward universally foundation models.
Reasoning Video Object Segmentation is a challenging task, which generates a mask sequence from an input video and an implicit, complex text query. Existing works probe into the problem by finetuning Multimodal Large Language Models (MLLM) for segmentation-based output, while still falling short in difficult cases on videos given temporally-sensitive queries, primarily due to the failure to integrate temporal and spatial information. In this paper, we propose ThinkVideo, a novel framework which leverages the zero-shot Chain-of-Thought (CoT) capability of MLLM to address these challenges. Specifically, ThinkVideo utilizes the CoT prompts to extract object selectivities associated with particular keyframes, then bridging the reasoning image segmentation model and SAM2 video processor to output mask sequences. The ThinkVideo framework is training-free and compatible with closed-source MLLMs, which can be applied to Reasoning Video Instance Segmentation. We further extend the framework for online video streams, where the CoT is used to update the object of interest when a better target starts to emerge and becomes visible. We conduct extensive experiments on video object segmentation with explicit and implicit queries. The results show that ThinkVideo significantly outperforms previous works in both cases, qualitatively and quantitatively.
Manual annotation of volumetric medical images, such as magnetic resonance imaging (MRI) and computed tomography (CT), is a labor-intensive and time-consuming process. Recent advancements in foundation models for video object segmentation, such as Segment Anything Model 2 (SAM 2), offer a potential opportunity to significantly speed up the annotation process by manually annotating one or a few slices and then propagating target masks across the entire volume. However, the performance of SAM 2 in this context varies. Our experiments show that relying on a single memory bank and attention module is prone to error propagation, particularly at boundary regions where the target is present in the previous slice but absent in the current one. To address this problem, we propose Short-Long Memory SAM 2 (SLM-SAM 2), a novel architecture that integrates distinct short-term and long-term memory banks with separate attention modules to improve segmentation accuracy. We evaluate SLM-SAM 2 on three public datasets covering organs, bones, and muscles across MRI and CT modalities. We show that the proposed method markedly outperforms the default SAM 2, achieving average Dice Similarity Coefficient improvement of 0.14 and 0.11 in the scenarios when 5 volumes and 1 volume are available for the initial adaptation, respectively. SLM-SAM 2 also exhibits stronger resistance to over-propagation, making a notable step toward more accurate automated annotation of medical images for segmentation model development.
Referring video object segmentation (RVOS) aims to segment objects in videos guided by natural language descriptions. We propose FS-RVOS, a Transformer-based model with two key components: a cross-modal affinity module and an instance sequence matching strategy, which extends FS-RVOS to multi-object segmentation (FS-RVMOS). Experiments show FS-RVOS and FS-RVMOS outperform state-of-the-art methods across diverse benchmarks, demonstrating superior robustness and accuracy.
Given a single labeled example, in-context segmentation aims to segment corresponding objects. This setting, known as one-shot segmentation in few-shot learning, explores the segmentation model's generalization ability and has been applied to various vision tasks, including scene understanding and image/video editing. While recent Segment Anything Models have achieved state-of-the-art results in interactive segmentation, these approaches are not directly applicable to in-context segmentation. In this work, we propose the Dual Consistency SAM (DC-SAM) method based on prompt-tuning to adapt SAM and SAM2 for in-context segmentation of both images and videos. Our key insights are to enhance the features of the SAM's prompt encoder in segmentation by providing high-quality visual prompts. When generating a mask prior, we fuse the SAM features to better align the prompt encoder. Then, we design a cycle-consistent cross-attention on fused features and initial visual prompts. Next, a dual-branch design is provided by using the discriminative positive and negative prompts in the prompt encoder. Furthermore, we design a simple mask-tube training strategy to adopt our proposed dual consistency method into the mask tube. Although the proposed DC-SAM is primarily designed for images, it can be seamlessly extended to the video domain with the support of SAM2. Given the absence of in-context segmentation in the video domain, we manually curate and construct the first benchmark from existing video segmentation datasets, named In-Context Video Object Segmentation (IC-VOS), to better assess the in-context capability of the model. Extensive experiments demonstrate that our method achieves 55.5 (+1.4) mIoU on COCO-20i, 73.0 (+1.1) mIoU on PASCAL-5i, and a J&F score of 71.52 on the proposed IC-VOS benchmark. Our source code and benchmark are available at https://github.com/zaplm/DC-SAM.
Complex video object segmentation continues to face significant challenges in small object recognition, occlusion handling, and dynamic scene modeling. This report presents our solution, which ranked second in the MOSE track of CVPR 2025 PVUW Challenge. Based on an existing segmentation framework, we propose an improved model named MASSeg for complex video object segmentation, and construct an enhanced dataset, MOSE+, which includes typical scenarios with occlusions, cluttered backgrounds, and small target instances. During training, we incorporate a combination of inter-frame consistent and inconsistent data augmentation strategies to improve robustness and generalization. During inference, we design a mask output scaling strategy to better adapt to varying object sizes and occlusion levels. As a result, MASSeg achieves a J score of 0.8250, F score of 0.9007, and a J&F score of 0.8628 on the MOSE test set.
Referring video object segmentation (RVOS) is a challenging task that requires the model to segment the object in a video given the language description. MeViS is a recently proposed dataset that contains motion expressions of the target objects, leading to a challenging benchmark, compared with existing RVOS benchmarks. On the other hand, for referring expression tasks, a new trend is to adopt multi-modal large language model (MLLM) to achieve better image and text alignment. In this report, we show that with a simple modification to the test time inference method on stronger MLLMs, we can lead to stronger results on MeVIS. In particular, we adopt the recent method Sa2VA, a unified model for dense grounded understanding of both images and videos. By enlarging the scope of key frames, without any further training, we can achieve the 3rd place in the 4th PVUW workshop.
Referring Video Object Segmentation (RVOS) aims to segment target objects throughout a video based on a text description. This task has attracted increasing attention in the field of computer vision due to its promising applications in video editing and human-agent interaction. Recently, ReferDINO has demonstrated promising performance in this task by adapting object-level vision-language knowledge from pretrained foundational image models. In this report, we further enhance its capabilities by incorporating the advantages of SAM2 in mask quality and object consistency. In addition, to effectively balance performance between single-object and multi-object scenarios, we introduce a conditional mask fusion strategy that adaptively fuses the masks from ReferDINO and SAM2. Our solution, termed ReferDINO-Plus, achieves 60.43 \(\mathcal{J}\&\mathcal{F}\) on MeViS test set, securing 2nd place in the MeViS PVUW challenge at CVPR 2025. The code is available at: https://github.com/iSEE-Laboratory/ReferDINO-Plus.
Video object segmentation is crucial for the efficient analysis of complex medical video data, yet it faces significant challenges in data availability and annotation. We introduce the task of one-shot medical video object segmentation, which requires separating foreground and background pixels throughout a video given only the mask annotation of the first frame. To address this problem, we propose a temporal contrastive memory network comprising image and mask encoders to learn feature representations, a temporal contrastive memory bank that aligns embeddings from adjacent frames while pushing apart distant ones to explicitly model inter-frame relationships and stores these features, and a decoder that fuses encoded image features and memory readouts for segmentation. We also collect a diverse, multi-source medical video dataset spanning various modalities and anatomies to benchmark this task. Extensive experiments demonstrate state-of-the-art performance in segmenting both seen and unseen structures from a single exemplar, showing ability to generalize from scarce labels. This highlights the potential to alleviate annotation burdens for medical video analysis. Code is available at https://github.com/MedAITech/TCMN.
Few-shot video object segmentation aims to reduce annotation costs; however, existing methods still require abundant dense frame annotations for training, which are scarce in the medical domain. We investigate an extremely low-data regime that utilizes annotations from only a few video frames and leverages existing labeled images to minimize costly video annotations. Specifically, we propose a two-phase framework. First, we learn a few-shot segmentation model using labeled images. Subsequently, to improve performance without full supervision, we introduce a spatiotemporal consistency relearning approach on medical videos that enforces consistency between consecutive frames. Constraints are also enforced between the image model and relearning model at both feature and prediction levels. Experiments demonstrate the superiority of our approach over state-of-the-art few-shot segmentation methods. Our model bridges the gap between abundant annotated medical images and scarce, sparsely labeled medical videos to achieve strong video segmentation performance in this low data regime. Code is available at https://github.com/MedAITech/RAB.
Referring video object segmentation aims to segment and track a target object in a video using a natural language prompt. Existing methods typically fuse visual and textual features in a highly entangled manner, processing multi-modal information together to generate per-frame masks. However, this approach often struggles with ambiguous target identification, particularly in scenes with multiple similar objects, and fails to ensure consistent mask propagation across frames. To address these limitations, we introduce FindTrack, a novel decoupled framework that separates target identification from mask propagation. FindTrack first adaptively selects a key frame by balancing segmentation confidence and vision-text alignment, establishing a robust reference for the target object. This reference is then utilized by a dedicated propagation module to track and segment the object across the entire video. By decoupling these processes, FindTrack effectively reduces ambiguities in target association and enhances segmentation consistency. We demonstrate that FindTrack outperforms existing methods on public benchmarks.
Active vision enables dynamic visual perception, offering an alternative to static feedforward architectures in computer vision, which rely on large datasets and high computational resources. Biological selective attention mechanisms allow agents to focus on salient Regions of Interest (ROIs), reducing computational demand while maintaining real-time responsiveness. Event-based cameras, inspired by the mammalian retina, enhance this capability by capturing asynchronous scene changes enabling efficient low-latency processing. To distinguish moving objects while the event-based camera is in motion the agent requires an object motion segmentation mechanism to accurately detect targets and center them in the visual field (fovea). Integrating event-based sensors with neuromorphic algorithms represents a paradigm shift, using Spiking Neural Networks to parallelize computation and adapt to dynamic environments. This work presents a Spiking Convolutional Neural Network bioinspired attention system for selective attention through object motion sensitivity. The system generates events via fixational eye movements using a Dynamic Vision Sensor integrated into the Speck neuromorphic hardware, mounted on a Pan-Tilt unit, to identify the ROI and saccade toward it. The system, characterized using ideal gratings and benchmarked against the Event Camera Motion Segmentation Dataset, reaches a mean IoU of 82.2% and a mean SSIM of 96% in multi-object motion segmentation. The detection of salient objects reaches 88.8% accuracy in office scenarios and 89.8% in low-light conditions on the Event-Assisted Low-Light Video Object Segmentation Dataset. A real-time demonstrator shows the system's 0.12 s response to dynamic scenes. Its learning-free design ensures robustness across perceptual scenes, making it a reliable foundation for real-time robotic applications serving as a basis for more complex architectures.
Referring video object segmentation (RVOS) aims to segment objects in a video according to textual descriptions, which requires the integration of multimodal information and temporal dynamics perception. The Segment Anything Model 2 (SAM 2) has shown great effectiveness across various video segmentation tasks. However, its application to offline RVOS is challenged by the translation of the text into effective prompts and a lack of global context awareness. In this paper, we propose a novel RVOS framework, termed MPG-SAM 2, to address these challenges. Specifically, MPG-SAM 2 employs a unified multimodal encoder to jointly encode video and textual features, generating semantically aligned video and text embeddings, along with multimodal class tokens. A mask prior generator utilizes the video embeddings and class tokens to create pseudo masks of target objects and global context. These masks are fed into the prompt encoder as dense prompts along with multimodal class tokens as sparse prompts to generate accurate prompts for SAM 2. To provide the online SAM 2 with a global view, we introduce a hierarchical global-historical aggregator, which allows SAM 2 to aggregate global and historical information of target objects at both pixel and object levels, enhancing the target representation and temporal consistency. Extensive experiments on several RVOS benchmarks demonstrate the superiority of MPG-SAM 2 and the effectiveness of our proposed modules.
In this paper, we address the challenges in unsupervised video object segmentation (UVOS) by proposing an efficient algorithm, termed MTNet, which concurrently exploits motion and temporal cues. Unlike previous methods that focus solely on integrating appearance with motion or on modeling temporal relations, our method combines both aspects by integrating them within a unified framework. MTNet is devised by effectively merging appearance and motion features during the feature extraction process within encoders, promoting a more complementary representation. To capture the intricate long-range contextual dynamics and information embedded within videos, a temporal transformer module is introduced, facilitating efficacious inter-frame interactions throughout a video clip. Furthermore, we employ a cascade of decoders all feature levels across all feature levels to optimally exploit the derived features, aiming to generate increasingly precise segmentation masks. As a result, MTNet provides a strong and compact framework that explores both temporal and cross-modality knowledge to robustly localize and track the primary object accurately in various challenging scenarios efficiently. Extensive experiments across diverse benchmarks conclusively show that our method not only attains state-of-the-art performance in unsupervised video object segmentation but also delivers competitive results in video salient object detection. These findings highlight the method's robust versatility and its adeptness in adapting to a range of segmentation tasks. Source code is available on https://github.com/hy0523/MTNet.
Referring video object segmentation aims to segment objects within a video corresponding to a given text description. Existing transformer-based temporal modeling approaches face challenges related to query inconsistency and the limited consideration of context. Query inconsistency produces unstable masks of different objects in the middle of the video. The limited consideration of context leads to the segmentation of incorrect objects by failing to adequately account for the relationship between the given text and instances. To address these issues, we propose the Multi-context Temporal Consistency Module (MTCM), which consists of an Aligner and a Multi-Context Enhancer (MCE). The Aligner removes noise from queries and aligns them to achieve query consistency. The MCE predicts text-relevant queries by considering multi-context. We applied MTCM to four different models, increasing performance across all of them, particularly achieving 47.6 J&F on the MeViS. Code is available at https://github.com/Choi58/MTCM.
This work presents Sa2VA, the first unified model for dense grounded understanding of both images and videos. Unlike existing multi-modal large language models, which are often limited to specific modalities and tasks, Sa2VA supports a wide range of image and video tasks, including referring segmentation and conversation, with minimal one-shot instruction tuning. Sa2VA combines SAM-2, a foundation video segmentation model, with LLaVA, an advanced vision-language model, and unifies text, image, and video into a shared LLM token space. Using the LLM, Sa2VA generates instruction tokens that guide SAM-2 in producing precise masks, enabling a grounded, multi-modal understanding of both static and dynamic visual content. Additionally, we introduce Ref-SAV, an auto-labeled dataset containing over 72k object expressions in complex video scenes, designed to boost model performance. We also manually validate 2k video objects in the Ref-SAV datasets to benchmark referring video object segmentation in complex environments. Experiments show that Sa2VA achieves state-of-the-art across multiple tasks, particularly in referring video object segmentation, highlighting its potential for complex real-world applications.
In this work, we focus on semi-supervised learning for video action detection. Video action detection requires spatiotemporal localization in addition to classification, and a limited amount of labels makes the model prone to unreliable predictions. We present Stable Mean Teacher, a simple end-to-end teacher-based framework that benefits from improved and temporally consistent pseudo labels. It relies on a novel Error Recovery (EoR) module, which learns from students' mistakes on labeled samples and transfers this knowledge to the teacher to improve pseudo labels for unlabeled samples. Moreover, existing spatiotemporal losses do not take temporal coherency into account and are prone to temporal inconsistencies. To address this, we present Difference of Pixels (DoP), a simple and novel constraint focused on temporal consistency, leading to coherent temporal detections. We evaluate our approach on four different spatiotemporal detection benchmarks: UCF101-24, JHMDB21, AVA, and YouTube-VOS. Our approach outperforms the supervised baselines for action detection by an average margin of 23.5% on UCF101-24, 16% on JHMDB21, and 3.3% on AVA. Using merely 10% and 20% of data, it provides competitive performance compared to the supervised baseline trained on 100% annotations on UCF101-24 and JHMDB21, respectively. We further evaluate its effectiveness on AVA for scaling to large-scale datasets and YouTube-VOS for video object segmentation, demonstrating its generalization capability to other tasks in the video domain. Code and models are publicly available.
Referring video object segmentation (RVOS) requires tracking and segmenting an object throughout a video according to a given natural language expression, demanding both complex motion understanding and the alignment of visual representations with language descriptions. Given these challenges, the recently proposed Segment Anything Model 2 (SAM2) emerges as a potential candidate due to its ability to generate coherent segmentation mask tracks across video frames, and provide an inherent spatio-temporal objectness in its object token representations. In this paper, we introduce SOLA (Selection by Object Language Alignment), a novel framework that leverages SAM2 object tokens as compact video-level object representations, which are aligned with language features through a lightweight track selection module. To effectively facilitate this alignment, we propose an IoU-based pseudo-labeling strategy, which bridges the modality gap between SAM2 representations with language features. Extensive experiments show that SOLA achieves state-of-the-art performance on the MeViS dataset and demonstrate that SOLA offers an effective solution for RVOS. Our project page is available at: https://cvlab-kaist.github.io/SOLA.
Referring Video Object Segmentation (RVOS) relies on natural language expressions to segment an object in a video clip. Existing methods restrict reasoning either to independent short clips, losing global context, or process the entire video offline, impairing their application in a streaming fashion. In this work, we aim to surpass these limitations and design an RVOS method capable of effectively operating in streaming-like scenarios while retaining contextual information from past frames. We build upon the Segment-Anything 2 (SAM2) model, that provides robust segmentation and tracking capabilities and is naturally suited for streaming processing. We make SAM2 wiser, by empowering it with natural language understanding and explicit temporal modeling at the feature extraction stage, without fine-tuning its weights, and without outsourcing modality interaction to external models. To this end, we introduce a novel adapter module that injects temporal information and multi-modal cues in the feature extraction process. We further reveal the phenomenon of tracking bias in SAM2 and propose a learnable module to adjust its tracking focus when the current frame features suggest a new object more aligned with the caption. Our proposed method, SAMWISE, achieves state-of-the-art across various benchmarks, by adding a negligible overhead of less than 5 M parameters. Code is available at https://github.com/ClaudiaCuttano/SAMWISE .