RSRefSeg 2: Decoupling Referring Remote Sensing Image Segmentation with Foundation Models

Referring Remote Sensing Image Segmentation provides a flexible and fine-grained framework for remote sensing scene analysis via vision-language collaborative interpretation. Current approaches predominantly utilize a three-stage pipeline encompassing dual-modal encoding, cross-modal interaction, and pixel decoding. These methods demonstrate significant limitations in managing complex semantic relationships and achieving precise cross-modal alignment, largely due to their coupled processing mechanism that conflates target localization with boundary delineation. This architectural coupling amplifies error propagation under semantic ambiguity while restricting model generalizability and interpretability. To address these issues, we propose RSRefSeg 2, a decoupling paradigm that reformulates the conventional workflow into a collaborative dual-stage framework: coarse localization followed by fine segmentation. RSRefSeg 2 integrates CLIP's cross-modal alignment strength with SAM's segmentation generalizability through strategic foundation model collaboration. Specifically, CLIP is employed as the dual-modal encoder to activate target features within its pre-aligned semantic space and generate localization prompts. To mitigate CLIP's misactivation challenges in multi-entity scenarios described by referring texts, a cascaded second-order prompter is devised, which enhances precision through implicit reasoning via decomposition of text embeddings into complementary semantic subspaces. These optimized semantic prompts subsequently direct the SAM to generate pixel-level refined masks, thereby completing the semantic transmission pipeline. Extensive experiments (RefSegRS, RRSIS-D, and RISBench) demonstrate that RSRefSeg 2 surpasses contemporary methods in segmentation accuracy (+~3% gIoU) and complex semantic interpretation. Code is available at: https://github.com/KyanChen/RSRefSeg2.

THOR: Thermal-guided Hand-Object Reasoning via Adaptive Vision Sampling

Wearable cameras are increasingly used as an observational and interventional tool for human behaviors by providing detailed visual data of hand-related activities. This data can be leveraged to facilitate memory recall for logging of behavior or timely interventions aimed at improving health. However, continuous processing of RGB images from these cameras consumes significant power impacting battery lifetime, generates a large volume of unnecessary video data for post-processing, raises privacy concerns, and requires substantial computational resources for real-time analysis. We introduce THOR, a real-time adaptive spatio-temporal RGB frame sampling method that leverages thermal sensing to capture hand-object patches and classify them in real-time. We use low-resolution thermal camera data to identify moments when a person switches from one hand-related activity to another, and adjust the RGB frame sampling rate by increasing it during activity transitions and reducing it during periods of sustained activity. Additionally, we use the thermal cues from the hand to localize the region of interest (i.e., the hand-object interaction) in each RGB frame, allowing the system to crop and process only the necessary part of the image for activity recognition. We develop a wearable device to validate our method through an in-the-wild study with 14 participants and over 30 activities, and further evaluate it on Ego4D (923 participants across 9 countries, totaling 3,670 hours of video). Our results show that using only 3% of the original RGB video data, our method captures all the activity segments, and achieves hand-related activity recognition F1-score (95%) comparable to using the entire RGB video (94%). Our work provides a more practical path for the longitudinal use of wearable cameras to monitor hand-related activities and health-risk behaviors in real time.

MUG: Pseudo Labeling Augmented Audio-Visual Mamba Network for Audio-Visual Video Parsing

The weakly-supervised audio-visual video parsing (AVVP) aims to predict all modality-specific events and locate their temporal boundaries. Despite significant progress, due to the limitations of the weakly-supervised and the deficiencies of the model architecture, existing methods are lacking in simultaneously improving both the segment-level prediction and the event-level prediction. In this work, we propose a audio-visual Mamba network with pseudo labeling aUGmentation (MUG) for emphasising the uniqueness of each segment and excluding the noise interference from the alternate modalities. Specifically, we annotate some of the pseudo-labels based on previous work. Using unimodal pseudo-labels, we perform cross-modal random combinations to generate new data, which can enhance the model's ability to parse various segment-level event combinations. For feature processing and interaction, we employ a audio-visual mamba network. The AV-Mamba enhances the ability to perceive different segments and excludes additional modal noise while sharing similar modal information. Our extensive experiments demonstrate that MUG improves state-of-the-art results on LLP dataset in all metrics (e.g,, gains of 2.1% and 1.2% in terms of visual Segment-level and audio Segment-level metrics). Our code is available at https://github.com/WangLY136/MUG.

Prompt Guidance and Human Proximal Perception for HOT Prediction with Regional Joint Loss

The task of Human-Object conTact (HOT) detection involves identifying the specific areas of the human body that are touching objects. Nevertheless, current models are restricted to just one type of image, often leading to too much segmentation in areas with little interaction, and struggling to maintain category consistency within specific regions. To tackle this issue, a HOT framework, termed \textbf{P3HOT}, is proposed, which blends \textbf{P}rompt guidance and human \textbf{P}roximal \textbf{P}erception. To begin with, we utilize a semantic-driven prompt mechanism to direct the network's attention towards the relevant regions based on the correlation between image and text. Then a human proximal perception mechanism is employed to dynamically perceive key depth range around the human, using learnable parameters to effectively eliminate regions where interactions are not expected. Calculating depth resolves the uncertainty of the overlap between humans and objects in a 2D perspective, providing a quasi-3D viewpoint. Moreover, a Regional Joint Loss (RJLoss) has been created as a new loss to inhibit abnormal categories in the same area. A new evaluation metric called ``AD-Acc.'' is introduced to address the shortcomings of existing methods in addressing negative samples. Comprehensive experimental results demonstrate that our approach achieves state-of-the-art performance in four metrics across two benchmark datasets. Specifically, our model achieves an improvement of \textbf{0.7}$\uparrow$, \textbf{2.0}$\uparrow$, \textbf{1.6}$\uparrow$, and \textbf{11.0}$\uparrow$ in SC-Acc., mIoU, wIoU, and AD-Acc. metrics, respectively, on the HOT-Annotated dataset. Code is available at https://github.com/YuxiaoWang-AI/P3HOT.

ViRefSAM: Visual Reference-Guided Segment Anything Model for Remote Sensing Segmentation

The Segment Anything Model (SAM), with its prompt-driven paradigm, exhibits strong generalization in generic segmentation tasks. However, applying SAM to remote sensing (RS) images still faces two major challenges. First, manually constructing precise prompts for each image (e.g., points or boxes) is labor-intensive and inefficient, especially in RS scenarios with dense small objects or spatially fragmented distributions. Second, SAM lacks domain adaptability, as it is pre-trained primarily on natural images and struggles to capture RS-specific semantics and spatial characteristics, especially when segmenting novel or unseen classes. To address these issues, inspired by few-shot learning, we propose ViRefSAM, a novel framework that guides SAM utilizing only a few annotated reference images that contain class-specific objects. Without requiring manual prompts, ViRefSAM enables automatic segmentation of class-consistent objects across RS images. Specifically, ViRefSAM introduces two key components while keeping SAM's original architecture intact: (1) a Visual Contextual Prompt Encoder that extracts class-specific semantic clues from reference images and generates object-aware prompts via contextual interaction with target images; and (2) a Dynamic Target Alignment Adapter, integrated into SAM's image encoder, which mitigates the domain gap by injecting class-specific semantics into target image features, enabling SAM to dynamically focus on task-relevant regions. Extensive experiments on three few-shot segmentation benchmarks, including iSAID-5$^i$, LoveDA-2$^i$, and COCO-20$^i$, demonstrate that ViRefSAM enables accurate and automatic segmentation of unseen classes by leveraging only a few reference images and consistently outperforms existing few-shot segmentation methods across diverse datasets.

Mamba Guided Boundary Prior Matters: A New Perspective for Generalized Polyp Segmentation

Polyp segmentation in colonoscopy images is crucial for early detection and diagnosis of colorectal cancer. However, this task remains a significant challenge due to the substantial variations in polyp shape, size, and color, as well as the high similarity between polyps and surrounding tissues, often compounded by indistinct boundaries. While existing encoder-decoder CNN and transformer-based approaches have shown promising results, they struggle with stable segmentation performance on polyps with weak or blurry boundaries. These methods exhibit limited abilities to distinguish between polyps and non-polyps and capture essential boundary cues. Moreover, their generalizability still falls short of meeting the demands of real-time clinical applications. To address these limitations, we propose SAM-MaGuP, a groundbreaking approach for robust polyp segmentation. By incorporating a boundary distillation module and a 1D-2D Mamba adapter within the Segment Anything Model (SAM), SAM-MaGuP excels at resolving weak boundary challenges and amplifies feature learning through enriched global contextual interactions. Extensive evaluations across five diverse datasets reveal that SAM-MaGuP outperforms state-of-the-art methods, achieving unmatched segmentation accuracy and robustness. Our key innovations, a Mamba-guided boundary prior and a 1D-2D Mamba block, set a new benchmark in the field, pushing the boundaries of polyp segmentation to new heights.

Are Vision Transformer Representations Semantically Meaningful? A Case Study in Medical Imaging

Vision transformers (ViTs) have rapidly gained prominence in medical imaging tasks such as disease classification, segmentation, and detection due to their superior accuracy compared to conventional deep learning models. However, due to their size and complex interactions via the self-attention mechanism, they are not well understood. In particular, it is unclear whether the representations produced by such models are semantically meaningful. In this paper, using a projected gradient-based algorithm, we show that their representations are not semantically meaningful and they are inherently vulnerable to small changes. Images with imperceptible differences can have very different representations; on the other hand, images that should belong to different semantic classes can have nearly identical representations. Such vulnerability can lead to unreliable classification results; for example, unnoticeable changes cause the classification accuracy to be reduced by over 60\%. %. To the best of our knowledge, this is the first work to systematically demonstrate this fundamental lack of semantic meaningfulness in ViT representations for medical image classification, revealing a critical challenge for their deployment in safety-critical systems.

DeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation through Loopback Synergy

Referring Image Segmentation (RIS) is a challenging task that aims to segment objects in an image based on natural language expressions. While prior studies have predominantly concentrated on improving vision-language interactions and achieving fine-grained localization, a systematic analysis of the fundamental bottlenecks in existing RIS frameworks remains underexplored. To bridge this gap, we propose DeRIS, a novel framework that decomposes RIS into two key components: perception and cognition. This modular decomposition facilitates a systematic analysis of the primary bottlenecks impeding RIS performance. Our findings reveal that the predominant limitation lies not in perceptual deficiencies, but in the insufficient multi-modal cognitive capacity of current models. To mitigate this, we propose a Loopback Synergy mechanism, which enhances the synergy between the perception and cognition modules, thereby enabling precise segmentation while simultaneously improving robust image-text comprehension. Additionally, we analyze and introduce a simple non-referent sample conversion data augmentation to address the long-tail distribution issue related to target existence judgement in general scenarios. Notably, DeRIS demonstrates inherent adaptability to both non- and multi-referents scenarios without requiring specialized architectural modifications, enhancing its general applicability. The codes and models are available at https://github.com/Dmmm1997/DeRIS.

TSDASeg: A Two-Stage Model with Direct Alignment for Interactive Point Cloud Segmentation

The rapid advancement of 3D vision-language models (VLMs) has spurred significant interest in interactive point cloud processing tasks, particularly for real-world applications. However, existing methods often underperform in point-level tasks, such as segmentation, due to missing direct 3D-text alignment, limiting their ability to link local 3D features with textual context. To solve this problem, we propose TSDASeg, a Two-Stage model coupled with a Direct cross-modal Alignment module and memory module for interactive point cloud Segmentation. We introduce the direct cross-modal alignment module to establish explicit alignment between 3D point clouds and textual/2D image data. Within the memory module, we employ multiple dedicated memory banks to separately store text features, visual features, and their cross-modal correspondence mappings. These memory banks are dynamically leveraged through self-attention and cross-attention mechanisms to update scene-specific features based on prior stored data, effectively addressing inconsistencies in interactive segmentation results across diverse scenarios. Experiments conducted on multiple 3D instruction, reference, and semantic segmentation datasets demonstrate that the proposed method achieves state-of-the-art performance.

Aligning Spoken Dialogue Models from User Interactions

We propose a novel preference alignment framework for improving spoken dialogue models on real-time conversations from user interactions. Current preference learning methods primarily focus on text-based language models, and are not directly suited to the complexities of real-time speech interactions, with richer dynamics (e.g. interruption, interjection) and no explicit segmentation between speaker turns.We create a large-scale dataset of more than 150,000 preference pairs from raw multi-turn speech conversations, annotated with AI feedback, to cover preferences over both linguistic content and temporal context variations. We leverage offline alignment methods to finetune a full-duplex autoregressive speech-to-speech model. Extensive experiments demonstrate that feedback on generic conversations can be consistently effective in improving spoken dialogue models to produce more factual, safer and more contextually aligned interactions. We deploy the finetuned model and conduct holistic human evaluations to assess the impact beyond single-turn conversations. Our findings shed light on the importance of a well-calibrated balance among various dynamics, crucial for natural real-time speech dialogue systems.