AAAI 2026 论文列表

We tackle the challenge of generating the infinitely extendable 3D world -- large, continuous environments with coherent geometry and realistic appearance. Existing methods face key challenges: 2D-lifting approaches suffer from geometric and appearance inconsistencies across views, 3D implicit representations are hard to scale up, and current 3D foundation models are mostly object-centric, limiting their applicability to scene-level generation. Our key insight is leveraging strong generation priors from pre-trained 3D models for structured scene block generation. To this end, we propose WorldGrow, a hierarchical framework for unbounded 3D scene synthesis. Our method features three core components: (1) a data curation pipeline that extracts high-quality scene blocks for training, making the 3D structured latent representations suitable for scene generation; (2) a 3D block inpainting mechanism that enables context-aware scene extension; and (3) a coarse-to-fine generation strategy that ensures both global layout plausibility and local geometric/textural fidelity. Evaluated on the large-scale 3D-FRONT dataset, WorldGrow achieves SOTA performance in geometry reconstruction, while uniquely supporting infinite scene generation with photorealistic and structurally consistent outputs. These results highlight its capability for constructing large-scale virtual environments and potential for building future world models.

Day-to-night unpaired image translation is important to downstream tasks but remains challenging due to large appearance shifts and the lack of direct pixel-level supervision. Existing methods often introduce semantic hallucinations, where objects from target classes such as traffic signs and vehicles, as well as man-made light effects, are incorrectly synthesized. These hallucinations significantly degrade downstream performance. We propose a novel framework that detects and suppresses hallucinations of target-class features during unpaired translation. To detect hallucination, we design a dual-head discriminator that additionly performs semantic segmentation to identify hallucinated content in background regions. To suppress these hallucinations, we introduce class-specific prototypes, constructed by aggregating features of annotated target-domain objects, which act as semantic anchors for each class. Built upon a Schrödinger Bridge-based translation model, our framework performs iterative refinement, where detected hallucination features are explicitly pushed away from class prototypes in feature space, thus preserving object semantics across the translation trajectory. Experiments show that our method outperforms existing approaches both qualitatively and quantitatively. On the BDD100K dataset, it improves mAP by 15.5% for day-to-night domain adaptation, with a notable 31.7% gain for classes such as traffic lights that are prone to hallucinations.

LiDAR-based 3D object detection models often struggle to generalize to real-world environments due to limited object diversity in existing datasets. To tackle it, we introduce the first generalized cross-domain few-shot (GCFS) task in 3D object detection, aiming to adapt a source-pretrained model to both common and novel classes in a new domain with only few-shot annotations. We propose a unified framework that learns stable target semantics under limited supervision by bridging 2D open-set semantics with 3D spatial reasoning. Specifically, an image-guided multi-modal fusion injects transferable 2D semantic cues into the 3D pipeline via vision-language models, while a physically-aware box search enhances 2D-to-3D alignment via LiDAR priors. To capture class-specific semantics from sparse data, we further introduce contrastive-enhanced prototype learning, which encodes few-shot instances into discriminative semantic anchors and stabilizes representation learning. Extensive experiments on GCFS benchmarks demonstrate the effectiveness and generality of our approach in realistic deployment settings.

The rapid advancement of generative models, which produce increasingly realistic synthetic images, urgently demands robust and generalizable detection methods. Consequently, research has largely pivoted to leveraging large-scale Vision Foundation Models (VFMs) for enhanced generalization. However, existing VFM-based approaches primarily adhere to either perceptual or generative paradigms, each with limitations: perceptual models capture high-level semantics but often miss subtle artifacts, whereas generative models emphasize fine-grained flaws yet overlook semantic inconsistency. To resolve this inherent trade-off, we introduce SynerDetect, a novel hierarchical synergistic framework that fundamentally unifies the two paradigms. SynerDetect achieves deep integration of heterogeneous forensic representations through two levels of synergy: Cross-Model Interactive Distillation (CMID) distills generative forensic signals into perceptual encoders via prompt-guided reconstruction; and Optimal Transport-Guided Discriminative Contrastive Learning (OT-DCL) structurally aligns and integrates these heterogeneous representations, consolidating them into a robust, unified detection space. SynerDetect achieves superior performance on standard benchmarks (AIGCDetectBenchmark and GenImage) and attains a notable 5.20% accuracy gain on the challenging Chameleon benchmark, whose synthetic images consistently pass the Visual Turing Test. These results unequivocally validate the robust, real-world generalization of our unified cross-paradigm framework.

Generating responsive listener head dynamics with nuanced emotions and expressive reactions is crucial for dialogue modeling in various virtual avatar animations. Previous studies mainly focus on the direct short-term production of listener behavior. They overlook the fine-grained control over motion variations and emotional intensity, especially in long-sequence modeling. Moreover, the lack of long-term and large-scale paired speaker-listener corpora incorporating head dynamics and fine-grained multi-modality annotations limits the application of dialogue modeling. Therefore, we first newly collect a large-scale multi-turn dataset of 3D dyadic conversation containing more than 1.4M valid frames for multi-modal responsive interaction, dubbed ListenerX. Additionally, we propose VividListener, a novel framework enabling fine-grained, expressive, and controllable listener dynamics modeling. This framework leverages multi-modal conditions as guiding principles for fostering coherent interactions between speakers and listeners. Specifically, we design the Responsive Interaction Module (RIM) to adaptively represent the multi-modal interactive embeddings. RIM ensures the listener dynamics achieve fine-grained semantic coordination with textual descriptions and adjustments, while preserving expressive reaction with speaker behavior. Meanwhile, we propose the Emotional Intensity Tags (EIT) for emotion intensity editing with multi-modal information integration, applying to both text descriptions and listener motion amplitude. Extensive experiments conducted on our newly collected ListenerX dataset demonstrate that VividListener achieves state-of-the-art performance, realizing expressive and controllable listener dynamics.

Accurate segmentation of ultra-high-resolution (UHR) images, which often exceed tens of millions of pixels, is critically important in domains such as remote sensing and biomedical imaging. However, acquiring pixel-level annotations for such high-resolution images is prohibitively expensive and labor-intensive. While semi-supervised semantic segmentation can significantly reduce the annotation burden, its extension to UHR images holds great potential for addressing the unique challenges posed by sparse supervision. To this end, we propose SSR-SAM, a retrieval-style semi-supervised segmentation framework tailored for UHR images. Leveraging the promptable paradigm of the Segment Anything Model (SAM), SSR-SAM treats locally annotated regions as prompts to retrieve semantically consistent pixels across the entire image. Building upon this retrieval-style segmentation paradigm, we further introduce prompt-level perturbation, a novel trail to deploy consistency regularization for semi-supervised segmentation. It encourages the model to learn consistency across predictions guided by diverse visual-semantic prompts, thereby enhancing generalization on unlabeled data. We evaluate SSR-SAM on three UHR datasets: Inria Aerial, BCSS, and URUR. Experimental results show that SSR-SAM achieves clear performance gains over the labeled-only supervision, with average mIoU improvements of 4.9%, 4.15%, and 2.5%, respectively. Additionally, SSR-SAM possesses zero-shot segmentation capability, exhibiting potential for general retrieval-style segmentation tasks.

Video Large Language Models (Video LLMs) have achieved significant success by adopting the paradigm of large-scale pre-training followed by supervised fine-tuning (SFT). However, existing approaches struggle with temporal reasoning due to weak temporal correspondence in the data and over-reliance on the next-token prediction paradigm, which collectively result in the absence temporal supervision. To address these limitations, we propose TEMPLE (TEMporal Preference Learning), a systematic framework that enhances temporal reasoning capabilities through Direct Preference Optimization (DPO). To address temporal information scarcity in data, we introduce an automated pipeline for systematically constructing temporality-intensive preference pairs comprising three steps: selecting temporally rich videos, designing video-specific perturbation strategies, and evaluating model responses on clean and perturbed inputs. Complementing this data pipeline, we provide additional supervision signals via preference learning and propose a novel Progressive Pre-SFT Alignment strategy featuring two key innovations: a curriculum learning strategy which progressively increases perturbation difficulty to maximize data efficiency; and applying preference optimization before instruction tuning to incentivize fundamental temporal alignment. Extensive experiments demonstrate that our approach consistently improves Video LLM performance across multiple benchmarks with a relatively small set of self-generated DPO data. Our findings highlight TEMPLE as a scalable and efficient complement to SFT-based methods, paving the way for developing reliable Video LLMs.

Despite significant advancements in image generation using advanced generative frameworks, cross-image integration of content and style remains a key challenge. Current generative models, while powerful, frequently depend on vague textual prompts to define styles—creating difficulties in balancing content semantics and style preservation. We propose a novel framework that utilizes customized models to learn style representations. It enhances content preservation through cross-model feature and attention modulation, leveraging the inherent semantic consistency across models. Additionally, we introduce fixed feature and adaptive attention fusion to achieve the desired balance between content and style. We further develop spatial (mask-guided localized) and temporal (multi-style compositional) multi-model combinations, enabling flexible fusion of models and styles. Extensive experiments demonstrate that our method outperforms state-of-the-art approaches in balancing content preservation and stylistic coherence.

LiDAR data generation has emerged as a promising solution to the high cost and limited scalability of real-world LiDAR sensing. Recent diffusion and rectified flow models have demonstrated strong capabilities in synthesizing realistic 3D point clouds; however, their iterative sampling procedures result in significant inference overhead. To address this, we focus on efficient few-step LiDAR generation for both unconditional and multi-modal conditional settings. Specifically, we propose an adaptive piecewise distillation strategy tailored for rectified flow-based LiDAR generation models, where the teacher model’s flow trajectory is adaptively segmented into consecutive intervals, and the student is trained only at the start of each interval to directly predict the velocity toward its endpoint. By sequentially sampling at the start timestep of each interval, our method enables fast few-step generation. Moreover, instead of uniform partitioning, we introduce an adaptive timestep selection strategy that chooses interval boundaries with minimal initial error, thereby reducing the complexity of distillation. Experimental results show that our method achieves comparable or superior performance to state-of-the-art methods in both unconditional and multi-modal conditional LiDAR generation, using only four sampling steps.

Deep functional map frameworks (DFM) for shape correspondence are powerful, yet fundamentally limited by their reliance on end-to-end differentiability. This constraint prevents the integration of highly accurate, non-differentiable refinement techniques, capping their overall performance, especially on challenging non-isometric shapes. To overcome this, we introduce MDND, a novel DFM paradigm built on the principle of merging differentiable and non-differentiable components. Our framework facilitates unsupervised learning guided by an internal, non-differentiable refinement. Specifically, MDND employs a dual-branch architecture: a non-differentiable refinement branch leverages a novel, multiscale iterative solver to produce highly robust correspondences, acting as a refined target. Concurrently, a fully differentiable branch learns to predict correspondences from features. The entire system is trained end-to-end without supervision by enforcing a consistency loss that compels the differentiable branch to learn from the superior, refined results of the non-differentiable branch. Extensive experiments show that MDND sets a new state-of-the-art, demonstrating remarkable robustness on shapes with non-isometric deformations and topological noise.

Vision-Language Models (VLMs) have demonstrated significant progress in quality assessment tasks. However, a fundamental paradox arises when their application to Point Cloud Quality Assessment (PCQA). Existing VLMs, designed for image-text pairs, are inherently incompatible with 3D point cloud data due to the modality gap. While some PCQA research attempts to adapt point clouds to VLMs by 2D projection, this approach inevitably sacrifices crucial spatial structure information essential for accurate quality assessment. Conversely, directly integrating a dedicated 3D branch into a VLM-based PCQA framework introduces feature space misalignment and an influx of quality-insensitive information. To bridge these fundamental conflicts hindering VLMs' adaptation to PCQA, we propose the PMP-PCQA framework, which leverages the inherent mapping relationship between points and pixels to seamlessly apply VLMs to PCQA. Our approach introduces three key innovations: a Spatial Awareness Enhancer(SAE) module that enriches the image features with spatial coordinate clues to reinforce geometric awareness in 2D visual representations; a Fine-to-coarse Consistency Alignment(FCA) module that bridges the gap between 2D and 3D modalities by leveraging point-pixel correspondences to construct bridging features; and a Text-Guided Adaptive Miner(TAM) module that dynamically suppresses quality-insensitive features to mine discriminative visual clues for PCQA. Extensive evaluations demonstrate that PMP-PCQA consistently outperforms state-of-the-art methods across multiple benchmarks.

Point Cloud Quality Assessment (PCQA) faces a critical disconnect: existing methods operate on a flawed single-perception paradigm, while human observers evaluate quality through dual cognitive streams: technical rationality and semantic sensibility. This fundamental mismatch routinely produces assessment failures in real-world scenarios where technical and semantic signals conflict. To address this, we introduce Dual-Stream Perception PCQA (DSP-PCQA), the first framework that explicitly models this perceptual duality through parallel networks thoroughly mirroring the human cognitive pathway. DSP-PCQA introduces three key innovations: (1) a Decoupled Focus Enhancer (DFE) that surgically isolates technical and semantic information using two targeted transformations; (2) a Context & Attribute Correlation Awareness (CACA) module that captures the dynamic, non-linear relationships between different views and sub-models characteristic of human visual processing; and (3) an Exchange-based Perceptual Injection (EPI) module that strategically transfers information between perception streams, simulating how humans integrate multiple perceptual dimensions. Extensive evaluations show DSP-PCQA outperforms state-of-the-art methods across multiple benchmarks. Most importantly, our method resolves the perceptual discord that plagues existing approaches, maintaining high accuracy even in the challenging boundary cases where technical quality and semantic significance diverge, precisely where conventional methods often struggle.

Deformable medical image registration is essential in medical image analyses. Recent transformer-based registration methods have achieved high registration accuracy. However, these methods often rely on patch embedding at the beginning of encoding, resulting in limited ability to capture detailed anatomical structural information in the images and explore local semantic relationships within individual patches. Here, we proposed a novel Dual-feet Encoder (DFEnc) to asynchronously model semantic information from moving and fixed images at various scales through two separate branches in three steps. For each step, features from adjacent resolution levels were processed by a Single Step Hybrid Extractor (SSHExt), which performed patch convolution to preserve local information, followed by several transformer blocks to capture global context. Dense connections were employed to enhance semantic awareness across adjacent feature resolution levels. Additionally, we introduced a Feature Fusion-based Decoder (FFDec) to progressively fuse features related to the fixed and moving images and to generate intermediate deformation fields at each stage, enabling accurate image alignment through stepwise warping and alignment refinement. Extensive ablation studies demonstrated the effectiveness of the proposed DFEnc, SSHExt, and FFDec. Compared to a state-of-the-art AutoFuse-Trans method, our approach yielded improvements in Dice of 1.14%, 1.77%, and 4.47% on the ACDC, OASIS, and Abdomen CT datasets, respectively, while maintaining relatively low computational cost. These results suggest the utility of the proposed approach for broad research and clinical applications.

Recent advances in generative AI have accelerated the production of ultra-high-resolution visual content. However, traditional image formats face significant limitations in efficient compression and real-time decoding, which restricts their applicability on end-user devices. Inspired by 3D Gaussian Splatting, 2D Gaussian image models have achieved notable progress in enhancing image representation efficiency and quality. Nevertheless, existing methods struggle to balance compression ratios and reconstruction fidelity in ultra-high-resolution scenarios. To address these challenges, we propose SmartSplat, a highly adaptive and feature-aware GS-based image compression framework that effectively supports arbitrary image resolutions and compression ratios. By leveraging image-aware features such as gradients and color variances, SmartSplat introduces a Gradient-Color Guided Variational Sampling strategy alongside an Exclusion-based Uniform Sampling scheme, significantly improving the non-overlapping coverage of Gaussian primitives in pixel space. Additionally, a Scale-Adaptive Gaussian Color Sampling method is proposed to enhance the initialization of Gaussian color attributes across scales. Through joint optimization of spatial layout, scale, and color initialization, SmartSplat can efficiently capture both local structures and global textures of images using a limited number of Gaussians, achieving superior reconstruction quality under high compression ratios. Extensive experiments on DIV8K and a newly created 16K dataset demonstrate that SmartSplat significantly outperforms state-of-the-art methods at comparable compression ratios and surpasses their compression limits, exhibiting strong scalability and practical applicability. This framework can effectively alleviate the storage and transmission burdens of ultra-high-resolution images, providing a robust foundation for future high-efficiency visual content processing.

Recent advances in multi-modal large language models (MLLMs) have significantly improved object-level grounding and region captioning. However, they remain limited in visual relation understanding, struggling even with binary relation detection, let alone N-ary relations involving multiple semantic roles. The core reason is the lack of modeling for structural semantic dependencies among multi-entities, leading to over-reliance on language priors (e.g., defaulting to "person drinks a milk" if a person is merely holding it). To this end, we propose Relation-R1, the first unified relation comprehension framework that explicitly integrates cognitive chain-of-thought (CoT)-guided supervised fine-tuning (SFT) and group relative policy optimization (GRPO) within a reinforcement learning (RL) paradigm. Specifically, we first establish foundational reasoning capabilities via SFT, enforcing structured outputs with thinking processes. Then, GRPO is utilized to refine these outputs via multi-rewards optimization, prioritizing visual-semantic grounding over language-induced biases, thereby improving generalization capability. Furthermore, we investigate the impact of various CoT strategies within this framework, demonstrating that a specific-to-general progressive approach in CoT guidance further improves generalization, especially in capturing synonymous N-ary relations. Extensive experiments on widely-used PSG and SWiG datasets demonstrate that Relation-R1 achieves state-of-the-art performance in both binary and N-ary relation understanding.

We introduce LLaMMo (Large Language and Multi-Person Motion Assistant), the first instruction-tuning multimodal framework tailored for multi-human motion analysis. LLaMMo incorporates a novel human-centric and social-temporal learner that models and fuses both intra-person dynamics and inter-person dependencies, yielding robust, context-aware representations of complex group behaviors while maintaining low computational overhead. To support LLaMMo, we construct LLaVerse, a large-scale dataset with fine-grained manual annotations covering diverse multi-person activities spanning daily social interaction and professional team sports. Built on top of LLaVerse, we also propose LLaMI-Bench, a dedicated benchmark for evaluating multi-human behavior understanding across motion and video modalities. Extensive experiments demonstrate that LLaMMo consistently outperforms baselines in understanding multi-person interactions under low-latency settings, with notable gains in both social and sport-specific contexts.

Remote sensing visual grounding (RSVG) aims to localize objects in remote sensing images based on free-form natural language expressions. Existing approaches are typically constrained to closed-set vocabularies, limiting their applicability in open-world scenarios. While recent attempts to leverage generic foundation models for open-vocabulary RSVG, they overly rely on expensive high-quality datasets and time-consuming fine-tuning. To address these limitations, we propose RSVG-ZeroOV, a training-free framework that aims to explore the potential of frozen generic foundation models for zero-shot open-vocabulary RSVG. Specifically, RSVG-ZeroOV comprises three key stages: (i) Overview: We utilize a vision-language model (VLM) to obtain cross-attention maps that capture semantic correlations between text queries and visual regions. (ii) Focus: By leveraging the fine-grained modeling priors of a diffusion model (DM), we fill in gaps in structural and shape information of objects, which are often overlooked by VLM. (iii) Evolve: A simple yet effective attention evolution module is introduced to suppress irrelevant activations, yielding purified segmentation masks over the referred objects. Without cumbersome task-specific training, RSVG-ZeroOV offers an efficient and scalable solution. Extensive experiments demonstrate that the proposed framework consistently outperforms existing weakly-supervised and zero-shot methods.

Monitoring Earth's evolving land covers requires methods capable of detecting changes across a wide range of categories and contexts. Existing change detection methods are hindered by their dependency on predefined classes, reducing their effectiveness in open-world applications. To address this issue, we introduce open-vocabulary change detection (OVCD), a novel task that bridges vision and language to detect changes across any category. Considering the lack of high-quality data and annotation, we propose two training-free frameworks, M-C-I and I-M-C, which leverage and integrate off-the-shelf foundation models for the OVCD task. The insight behind the M-C-I~framework is to discover all potential changes and then classify these changes, while the insight of I-M-C~framework is to identify all targets of interest and then determine whether their states have changed. Based on these two frameworks, we instantiate to obtain several methods, e.g., SAM-DINOv2-SegEarth-OV, Grounding-DINO-SAM2-DINO, etc. Extensive evaluations on 4 benchmark datasets demonstrate the superior generalization and robustness of our OVCD methods over existing supervised and unsupervised methods. To support continued exploration, we release DynamicEarth, a dedicated codebase designed to advance research and application of OVCD.

The widespread and inconsistent compression applied by Online Social Networks severely degrades the performance of synthetic image detectors. We attribute this degradation to two main issues: 1) the model confuses forgery artifacts with compression artifacts, and 2) compression erodes crucial discriminative high-frequency details. Existing methods suppress compression features during training but overlook the overlap between compression features and forgery-related features, leading to the unintended removal of forgery traces. To address artifact confusion, we introduce a Decision-Driven Orthogonal Constraint, which defines a classification decision axis pointing from the real class centroid to the forged class centroid. This constraint enforces compression artifacts to be orthogonal to the decision axis, mitigating their interference with forgery detection without entirely removing them, thus preventing the suppression of forgery-related features. To mitigate the erosion of high-frequency details, we propose to mine complementary forgery cues from both low-frequency information and compressed high-frequency components. A bidirectional update strategy and an adaptive global-local modulator are proposed to facilitate the utilization of forgery cues. Extensive experiments demonstrate that our method achieves state-of-the-art generalization performance in challenging open-world detection scenarios.