Tianrun Chen

Xiao Fu, Shangzhan Zhang, Tianrun Chen et al.

Training perception systems for self-driving cars requires substantial 2D annotations that are labor-intensive to manual label. While existing datasets provide rich annotations on pre-recorded sequences, they fall short in labeling rarely encountered viewpoints, potentially hampering the generalization ability for perception models. In this paper, we present PanopticNeRF-360, a novel approach that combines coarse 3D annotations with noisy 2D semantic cues to generate high-quality panoptic labels and images from any viewpoint. Our key insight lies in exploiting the complementarity of 3D and 2D priors to mutually enhance geometry and semantics. Specifically, we propose to leverage coarse 3D bounding primitives and noisy 2D semantic and instance predictions to guide geometry optimization, by encouraging predicted labels to match panoptic pseudo ground truth. Simultaneously, the improved geometry assists in filtering 3D&2D annotation noise by fusing semantics in 3D space via a learned semantic field. To further enhance appearance, we combine MLP and hash grids to yield hybrid scene features, striking a balance between high-frequency appearance and contiguous semantics. Our experiments demonstrate PanopticNeRF-360's state-of-the-art performance over label transfer methods on the challenging urban scenes of the KITTI-360 dataset. Moreover, PanopticNeRF-360 enables omnidirectional rendering of high-fidelity, multi-view and spatiotemporally consistent appearance, semantic and instance labels. We make our code and data available at https://github.com/fuxiao0719/PanopticNeRF

Xiao Fu, Shangzhan Zhang, Tianrun Chen et al.

Large-scale training data with high-quality annotations is critical for training semantic and instance segmentation models. Unfortunately, pixel-wise annotation is labor-intensive and costly, raising the demand for more efficient labeling strategies. In this work, we present a novel 3D-to-2D label transfer method, Panoptic NeRF, which aims for obtaining per-pixel 2D semantic and instance labels from easy-to-obtain coarse 3D bounding primitives. Our method utilizes NeRF as a differentiable tool to unify coarse 3D annotations and 2D semantic cues transferred from existing datasets. We demonstrate that this combination allows for improved geometry guided by semantic information, enabling rendering of accurate semantic maps across multiple views. Furthermore, this fusion process resolves label ambiguity of the coarse 3D annotations and filters noise in the 2D predictions. By inferring in 3D space and rendering to 2D labels, our 2D semantic and instance labels are multi-view consistent by design. Experimental results show that Panoptic NeRF outperforms existing label transfer methods in terms of accuracy and multi-view consistency on challenging urban scenes of the KITTI-360 dataset.

Jiatong Li, Zheng Chen, Kai Liu et al.

This paper provides a review of the NTIRE 2026 challenge on mobile real-world image super-resolution, highlighting the proposed solutions and the resulting outcomes. The challenge aims to recover high-resolution (HR) images from low-resolution (LR) counterparts generated through unknown degradations with a x4 scaling factor while ensuring the models remain executable on mobile devices. The objective is to develop effective and efficient network designs or solutions that achieve state-of-the-art real-world image super-resolution performance. The track of the challenge evaluates performance using a weighted combination of image quality assessment (IQA) score and speedup ratios. The competition attracted 108 registrants, with 16 teams achieving a valid score in the final ranking. This collaborative effort advances the performance of mobile real-world image super-resolution while offering an in-depth overview of the latest trends in the field.

Lanyun Zhu, Tianrun Chen, Jianxiong Yin et al.

Extracting and using class-discriminative features is critical for fine-grained recognition. Existing works have demonstrated the possibility of applying deep CNNs to exploit features that distinguish similar classes. However, CNNs suffer from problems including frequency bias and loss of detailed local information, which restricts the performance of recognizing fine-grained categories. To address the challenge, we propose a novel texture branch as complimentary to the CNN branch for feature extraction. We innovatively utilize Gabor filters as a powerful extractor to exploit texture features, motivated by the capability of Gabor filters in effectively capturing multi-frequency features and detailed local information. We implement several designs to enhance the effectiveness of Gabor filters, including imposing constraints on parameter values and developing a learning method to determine the optimal parameters. Moreover, we introduce a statistical feature extractor to utilize informative statistical information from the signals captured by Gabor filters, and a gate selection mechanism to enable efficient computation by only considering qualified regions as input for texture extraction. Through the integration of features from the Gabor-filter-based texture branch and CNN-based semantic branch, we achieve comprehensive information extraction. We demonstrate the efficacy of our method on multiple datasets, including CUB-200-2011, NA-bird, Stanford Dogs, and GTOS-mobile. State-of-the-art performance is achieved using our approach.

Lanyun Zhu, Tianrun Chen, Jianxiong Yin et al.

Continual Semantic Segmentation (CSS) extends static semantic segmentation by incrementally introducing new classes for training. To alleviate the catastrophic forgetting issue in CSS, a memory buffer that stores a small number of samples from the previous classes is constructed for replay. However, existing methods select the memory samples either randomly or based on a single-factor-driven handcrafted strategy, which has no guarantee to be optimal. In this work, we propose a novel memory sample selection mechanism that selects informative samples for effective replay in a fully automatic way by considering comprehensive factors including sample diversity and class performance. Our mechanism regards the selection operation as a decision-making process and learns an optimal selection policy that directly maximizes the validation performance on a reward set. To facilitate the selection decision, we design a novel state representation and a dual-stage action space. Our extensive experiments on Pascal-VOC 2012 and ADE 20K datasets demonstrate the effectiveness of our approach with state-of-the-art (SOTA) performance achieved, outperforming the second-place one by 12.54% for the 6stage setting on Pascal-VOC 2012.

Zheng Chen, Kai Liu, Jingkai Wang et al.

This paper presents the NTIRE 2026 image super-resolution ($\times$4) challenge, one of the associated competitions of the NTIRE 2026 Workshop at CVPR 2026. The challenge aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs generated through bicubic downsampling with a $\times$4 scaling factor. The objective is to develop effective super-resolution solutions and analyze recent advances in the field. To reflect the evolving objectives of image super-resolution, the challenge includes two tracks: (1) a restoration track, which emphasizes pixel-wise fidelity and ranks submissions based on PSNR; and (2) a perceptual track, which focuses on visual realism and evaluates results using a perceptual score. A total of 194 participants registered for the challenge, with 31 teams submitting valid entries. This report summarizes the challenge design, datasets, evaluation protocol, main results, and methods of participating teams. The challenge provides a unified benchmark and offers insights into current progress and future directions in image super-resolution.

Shangzan Zhang, Sida Peng, Yinji ShenTu et al.

Recently, the editing of neural radiance fields (NeRFs) has gained considerable attention, but most prior works focus on static scenes while research on the appearance editing of dynamic scenes is relatively lacking. In this paper, we propose a novel framework to edit the local appearance of dynamic NeRFs by manipulating pixels in a single frame of training video. Specifically, to locally edit the appearance of dynamic NeRFs while preserving unedited regions, we introduce a local surface representation of the edited region, which can be inserted into and rendered along with the original NeRF and warped to arbitrary other frames through a learned invertible motion representation network. By employing our method, users without professional expertise can easily add desired content to the appearance of a dynamic scene. We extensively evaluate our approach on various scenes and show that our approach achieves spatially and temporally consistent editing results. Notably, our approach is versatile and applicable to different variants of dynamic NeRF representations.

Lanyun Zhu, Tianrun Chen, Deyi Ji et al.

This paper proposes LLaFS, the first attempt to leverage large language models (LLMs) in few-shot segmentation. In contrast to the conventional few-shot segmentation methods that only rely on the limited and biased information from the annotated support images, LLaFS leverages the vast prior knowledge gained by LLM as an effective supplement and directly uses the LLM to segment images in a few-shot manner. To enable the text-based LLM to handle image-related tasks, we carefully design an input instruction that allows the LLM to produce segmentation results represented as polygons, and propose a region-attribute table to simulate the human visual mechanism and provide multi-modal guidance. We also synthesize pseudo samples and use curriculum learning for pretraining to augment data and achieve better optimization. LLaFS achieves state-of-the-art results on multiple datasets, showing the potential of using LLMs for few-shot computer vision tasks.

Shangzan Zhang, Sida Peng, Tianrun Chen et al.

We introduce a novel approach that takes a single semantic mask as input to synthesize multi-view consistent color images of natural scenes, trained with a collection of single images from the Internet. Prior works on 3D-aware image synthesis either require multi-view supervision or learning category-level prior for specific classes of objects, which can hardly work for natural scenes. Our key idea to solve this challenging problem is to use a semantic field as the intermediate representation, which is easier to reconstruct from an input semantic mask and then translate to a radiance field with the assistance of off-the-shelf semantic image synthesis models. Experiments show that our method outperforms baseline methods and produces photorealistic, multi-view consistent videos of a variety of natural scenes.

Tianrun Chen, Lanyun Zhu, Chaotao Ding et al.

The emergence of large models, also known as foundation models, has brought significant advancements to AI research. One such model is Segment Anything (SAM), which is designed for image segmentation tasks. However, as with other foundation models, our experimental findings suggest that SAM may fail or perform poorly in certain segmentation tasks, such as shadow detection and camouflaged object detection (concealed object detection). This study first paves the way for applying the large pre-trained image segmentation model SAM to these downstream tasks, even in situations where SAM performs poorly. Rather than fine-tuning the SAM network, we propose \textbf{SAM-Adapter}, which incorporates domain-specific information or visual prompts into the segmentation network by using simple yet effective adapters. By integrating task-specific knowledge with general knowledge learnt by the large model, SAM-Adapter can significantly elevate the performance of SAM in challenging tasks as shown in extensive experiments. We can even outperform task-specific network models and achieve state-of-the-art performance in the task we tested: camouflaged object detection, shadow detection. We also tested polyp segmentation (medical image segmentation) and achieves better results. We believe our work opens up opportunities for utilizing SAM in downstream tasks, with potential applications in various fields, including medical image processing, agriculture, remote sensing, and more.

Tianrun Chen, Ankang Lu, Lanyun Zhu et al.

The advent of large models, also known as foundation models, has significantly transformed the AI research landscape, with models like Segment Anything (SAM) achieving notable success in diverse image segmentation scenarios. Despite its advancements, SAM encountered limitations in handling some complex low-level segmentation tasks like camouflaged object and medical imaging. In response, in 2023, we introduced SAM-Adapter, which demonstrated improved performance on these challenging tasks. Now, with the release of Segment Anything 2 (SAM2), a successor with enhanced architecture and a larger training corpus, we reassess these challenges. This paper introduces SAM2-Adapter, the first adapter designed to overcome the persistent limitations observed in SAM2 and achieve new state-of-the-art (SOTA) results in specific downstream tasks including medical image segmentation, camouflaged (concealed) object detection, and shadow detection. SAM2-Adapter builds on the SAM-Adapter's strengths, offering enhanced generalizability and composability for diverse applications. We present extensive experimental results demonstrating SAM2-Adapter's effectiveness. We show the potential and encourage the research community to leverage the SAM2 model with our SAM2-Adapter for achieving superior segmentation outcomes. Code, pre-trained models, and data processing protocols are available at http://tianrun-chen.github.io/SAM-Adaptor/

Tianrun Chen, Chenglong Fu, Ying Zang et al.

The rapid development of AR/VR brings tremendous demands for 3D content. While the widely-used Computer-Aided Design (CAD) method requires a time-consuming and labor-intensive modeling process, sketch-based 3D modeling offers a potential solution as a natural form of computer-human interaction. However, the sparsity and ambiguity of sketches make it challenging to generate high-fidelity content reflecting creators' ideas. Precise drawing from multiple views or strategic step-by-step drawings is often required to tackle the challenge but is not friendly to novice users. In this work, we introduce a novel end-to-end approach, Deep3DSketch+, which performs 3D modeling using only a single free-hand sketch without inputting multiple sketches or view information. Specifically, we introduce a lightweight generation network for efficient inference in real-time and a structural-aware adversarial training approach with a Stroke Enhancement Module (SEM) to capture the structural information to facilitate learning of the realistic and fine-detailed shape structures for high-fidelity performance. Extensive experiments demonstrated the effectiveness of our approach with the state-of-the-art (SOTA) performance on both synthetic and real datasets.

Runlong Cao, Ying Zang, Chuanwei Zhou et al.

Semi-supervised referring expression segmentation (SS-RES) aims to achieve precise pixel-level language grounding under limited annotation, yet suffers from limited supervision and unreliable pseudo-labels when exploiting unlabeled image-text pairs. In this work, we propose Learning to Label, a reinforced self-evolving framework (L2L) that casts pseudo-label construction as a learnable decision-making process. To build foundational understanding, we leverage a multimodal large language model to extract semantic-spatial priors, which are instantiated as initial soft segmentation proposals and elevated, together with textual cues, into learnable guidance signals that condition a hierarchical segmentation network. To ensure stable learning, reinforced pseudo-label selection is formulated as an exploratory decision process that adaptively rewards high-utility pixel-level supervision based on multimodal priors and model predictions. This reinforced self-evolving loop enables joint optimization of the segmentation model and pseudo-labels, progressively enhancing label reliability under sparse supervision. Extensive experiments on RefCOCO, RefCOCO+, and RefCOCOg demonstrate improvements over existing methods, validating its effectiveness and generalization.

Tianrun Chen, Chaotao Ding, Lanyun Zhu et al.

Convolutional Neural Networks (CNNs) and Vision Transformers (ViT) have been pivotal in biomedical image segmentation, yet their ability to manage long-range dependencies remains constrained by inherent locality and computational overhead. To overcome these challenges, in this technical report, we first propose xLSTM-UNet, a UNet structured deep learning neural network that leverages Vision-LSTM (xLSTM) as its backbone for medical image segmentation. xLSTM is a recently proposed as the successor of Long Short-Term Memory (LSTM) networks and have demonstrated superior performance compared to Transformers and State Space Models (SSMs) like Mamba in Neural Language Processing (NLP) and image classification (as demonstrated in Vision-LSTM, or ViL implementation). Here, xLSTM-UNet we designed extend the success in biomedical image segmentation domain. By integrating the local feature extraction strengths of convolutional layers with the long-range dependency capturing abilities of xLSTM, xLSTM-UNet offers a robust solution for comprehensive image analysis. We validate the efficacy of xLSTM-UNet through experiments. Our findings demonstrate that xLSTM-UNet consistently surpasses the performance of leading CNN-based, Transformer-based, and Mamba-based segmentation networks in multiple datasets in biomedical segmentation including organs in abdomen MRI, instruments in endoscopic images, and cells in microscopic images. With comprehensive experiments performed, this technical report highlights the potential of xLSTM-based architectures in advancing biomedical image analysis in both 2D and 3D. The code, models, and datasets are publicly available at http://tianrun-chen.github.io/xLSTM-UNet/

Xuanyi Liu, Chunan Yu, Deyi Ji et al.

Reconstructing dense 3D geometry from continuous video streams requires stable inference under a constant memory budget. Existing $O(1)$ frameworks primarily rely on a ``pure eviction'' paradigm, which suffers from significant information destruction due to binary token deletion and evaluation noise from localized, single-layer scoring. To address these bottlenecks, we propose StreamCacheVGGT, a training-free framework that reimagines cache management through two synergistic modules: Cross-Layer Consistency-Enhanced Scoring (CLCES) and Hybrid Cache Compression (HCC). CLCES mitigates activation noise by tracking token importance trajectories across the Transformer hierarchy, employing order-statistical analysis to identify sustained geometric salience. Leveraging these robust scores, HCC transcends simple eviction by introducing a three-tier triage strategy that merges moderately important tokens into retained anchors via nearest-neighbor assignment on the key-vector manifold. This approach preserves essential geometric context that would otherwise be lost. Extensive evaluations on five benchmarks (7-Scenes, NRGBD, ETH3D, Bonn, and KITTI) demonstrate that StreamCacheVGGT sets a new state-of-the-art, delivering superior reconstruction accuracy and long-term stability while strictly adhering to constant-cost constraints.

Tianrun Chen, Yuanqi Hu, Yidong Han et al.

High-resolution imagery is essential for accurate 3D reconstruction, as many geometric details only emerge at fine spatial scales. Recent feed-forward approaches, such as the Visual Geometry Grounded Transformer (VGGT), have demonstrated the ability to infer scene geometry from large collections of images in a single forward pass. However, scaling these models to high-resolution inputs remains challenging: the number of tokens in transformer architectures grows rapidly with both image resolution and the number of views, leading to prohibitive computational and memory costs. Moreover, we observe that visually ambiguous regions, such as repetitive patterns, weak textures, or specular surfaces, often produce unstable feature tokens that degrade geometric inference, especially at higher resolutions. We introduce HD-VGGT, a dual-branch architecture for efficient and robust high-resolution 3D reconstruction. A low-resolution branch predicts a coarse, globally consistent geometry, while a high-resolution branch refines details via a learned feature upsampling module. To handle unstable tokens, we propose Feature Modulation, which suppresses unreliable features early in the transformer. HD-VGGT leverages high-resolution images and supervision without full-resolution transformer costs, achieving state-of-the-art reconstruction quality.

Fan Wu, Cheng Chen, Zhenshan Tan et al.

We present Claw AI Lab, a lab-native autonomous research platform that advances automated research from a hidden prompt-to-paper pipeline into an interactive AI laboratory. Rather than centering the system around a single agent or a fixed serial workflow, we allow users to instantiate a full research team from one prompt, with customizable roles, collaborative workflows, real-time monitoring, artifact inspection, and rollback/resume control through a unified dashboard. The platform also supports distinct research modes for exploration, multi-agent discussion, and reproduction, making autonomous research substantially more steerable and laboratory-like in practice. A key practical contribution of Claw AI Lab lies in its Claw-Code Harness, which connects local codebases, datasets, and checkpoints to runnable experiments and feeds execution artifacts back into the research loop. As a result, the harness improves not only execution integration, but also experimental completion and result integrity: experiments are easier to inspect, iterate on, and faithfully transfer into final papers, reducing common failure modes such as partial runs and malformed result reporting. In our internal evaluation on five AI research case studies, using AutoResearchClaw as the baseline, Claw AI Lab is consistently preferred by AI expert judges on idea novelty, experiment completeness, and paper presentation quality. We view Claw AI Lab as an early step toward a new paradigm: autonomous research as usable, interactive, and reliability-aware scientific infrastructure.

Ying Zang, Yidong Han, Chaotao Ding et al.

The requirement for 3D content is growing as AR/VR application emerges. At the same time, 3D modelling is only available for skillful experts, because traditional methods like Computer-Aided Design (CAD) are often too labor-intensive and skill-demanding, making it challenging for novice users. Our proposed method, Magic3DSketch, employs a novel technique that encodes sketches to predict a 3D mesh, guided by text descriptions and leveraging external prior knowledge obtained through text and language-image pre-training. The integration of language-image pre-trained neural networks complements the sparse and ambiguous nature of single-view sketch inputs. Our method is also more useful and offers higher degree of controllability compared to existing text-to-3D approaches, according to our user study. Moreover, Magic3DSketch achieves state-of-the-art performance in both synthetic and real dataset with the capability of producing more detailed structures and realistic shapes with the help of text input. Users are also more satisfied with models obtained by Magic3DSketch according to our user study. Additionally, we are also the first, to our knowledge, add color based on text description to the sketch-derived shapes. By combining sketches and text guidance with the help of language-image pretrained models, our Magic3DSketch can allow novice users to create custom 3D models with minimal effort and maximum creative freedom, with the potential to revolutionize future 3D modeling pipelines.

Zhenshan Tan, Chenhan Lu, Yuxiang Huang et al.

Image Manipulation Localization (IML) aims to identify edited regions in an image. However, with the increasing use of modern image editing and generative models, many manipulations no longer exhibit obvious low-level artifacts. Instead, they often involve subtle but meaning-altering edits to an object's attributes, state, or relationships while remaining highly consistent with the surrounding content. This makes conventional IML methods less effective because they mainly rely on artifact detection rather than semantic sensitivity. To address this issue, we introduce Semantic Manipulation Localization (SML), a new task that focuses on localizing subtle semantic edits that significantly change image interpretation. We further construct a dedicated fine-grained benchmark for SML using a semantics-driven manipulation pipeline with pixel-level annotations. Based on this task, we propose TRACE (Targeted Reasoning of Attributed Cognitive Edits), an end-to-end framework that models semantic sensitivity through three progressively coupled components: semantic anchoring, semantic perturbation sensing, and semantic-constrained reasoning. Specifically, TRACE first identifies semantically meaningful regions that support image understanding, then injects perturbation-sensitive frequency cues to capture subtle edits under strong visual consistency, and finally verifies candidate regions through joint reasoning over semantic content and semantic scope. Extensive experiments show that TRACE consistently outperforms existing IML methods on our benchmark and produces more complete, compact, and semantically coherent localization results. These results demonstrate the necessity of moving beyond artifact-based localization and provide a new direction for image forensics in complex semantic editing scenarios.

Chaojun Ni, Cheng Chen, Xiaofeng Wang et al.

Vision-Language-Action (VLA) models built on pretrained Vision-Language Models (VLMs) show strong potential but are limited in practicality due to their large parameter counts. To mitigate this issue, using a lightweight VLM has been explored, but it compromises spatiotemporal reasoning. Although some methods suggest that incorporating additional 3D inputs can help, they usually rely on large VLMs to fuse 3D and 2D inputs and still lack temporal understanding. Therefore, we propose SwiftVLA, an architecture that enhances a compact model with 4D understanding while preserving design efficiency. Specifically, our approach features a pretrained 4D visual geometry transformer with a temporal cache that extracts 4D features from 2D images. Then, to enhance the VLM's ability to exploit both 2D images and 4D features, we introduce Fusion Tokens, a set of learnable tokens trained with a future prediction objective to generate unified representations for action generation. Finally, we introduce a mask-and-reconstruct strategy that masks 4D inputs to the VLM and trains the VLA to reconstruct them, enabling the VLM to learn effective 4D representations and allowing the 4D branch to be dropped at inference with minimal performance loss. Experiments in real and simulated environments show that SwiftVLA outperforms lightweight baselines and rivals VLAs up to 7 times larger, achieving comparable performance on edge devices while being 18 times faster and reducing memory footprint by 12 times.

Shengyuan Zhang, An Zhao, Ling Yang et al.

Diffusion models have been applied to 3D LiDAR scene completion due to their strong training stability and high completion quality. However, the slow sampling speed limits the practical application of diffusion-based scene completion models since autonomous vehicles require an efficient perception of surrounding environments. This paper proposes a novel distillation method tailored for 3D Li- DAR scene completion models, dubbed ScoreLiDAR, which achieves efficient yet high-quality scene completion. Score- LiDAR enables the distilled model to sample in significantly fewer steps after distillation. To improve completion quality, we also introduce a novel Structural Loss, which encourages the distilled model to capture the geometric structure of the 3D LiDAR scene. The loss contains a scene-wise term constraining the holistic structure and a point-wise term constraining the key landmark points and their relative configuration. Extensive experiments demonstrate that ScoreLiDAR significantly accelerates the completion time from 30.55 to 5.37 seconds per frame (>5x) on SemanticKITTI and achieves superior performance compared to state-of-the-art 3D LiDAR scene completion models. Our model and code are publicly available on https://github.com/happyw1nd/ScoreLiDAR.

Ying Zang, Runlong Cao, Jianqi Zhang et al.

Sketches, with their expressive potential, allow humans to convey the essence of an object through even a rough contour. For the first time, we harness this expressive potential to improve segmentation performance in challenging tasks like camouflaged object detection (COD). Our approach introduces an innovative sketch-guided interactive segmentation framework, allowing users to intuitively annotate objects with freehand sketches (drawing a rough contour of the object) instead of the traditional bounding boxes or points used in classic interactive segmentation models like SAM. We demonstrate that sketch input can significantly improve performance in existing iterative segmentation methods, outperforming text or bounding box annotations. Additionally, we introduce key modifications to network architectures and a novel sketch augmentation technique to fully harness the power of sketch input and further boost segmentation accuracy. Remarkably, our model' s output can be directly used to train other neural networks, achieving results comparable to pixel-by-pixel annotations--while reducing annotation time by up to 120 times, which shows great potential in democratizing the annotation process and enabling model training with less reliance on resource-intensive, laborious pixel-level annotations. We also present KOSCamo+, the first freehand sketch dataset for camouflaged object detection. The dataset, code, and the labeling tool will be open sourced.

Ying Zang, Xuanyi Liu, Yidong Han et al.

Reconstructing dynamic 4D scenes from monocular videos is a fundamental yet challenging task. While recent 3D foundation models provide strong geometric priors, their performance significantly degrades in dynamic environments. This degradation stems from a fundamental tension: the inherent coupling of camera ego-motion and object motion within global attention mechanisms. In this paper, we propose a novel, training-free progressive decoupling framework that disentangles dynamics from statics in a principled, coarse-to-fine manner. Our core insight is to resolve the tension by first stabilizing the camera pose, followed by geometric refinement. Specifically, our approach consists of three synergistic components: (1) a Dynamic-Mask-Guided Pose Decoupling module that isolates pose estimation from dynamic interference, yielding a stable motion-free reference frame; (2) a Topological Subspace Surgery mechanism that orthogonally decomposes the depth manifold, safely preserving dynamic objects while injecting refined, mask-aware geometry into static regions; and (3) an Information-Theoretic Confidence-Aware Fusion strategy that formulates depth integration as a heteroscedastic Bayesian inference problem, adaptively blending multi-pass predictions via inverse-variance weighting. Extensive experiments on standard 4D reconstruction benchmarks demonstrate that our method achieves consistent and substantial improvements across principal point-cloud metrics. Notably, our approach shows competitive performance in robust 4D scene reconstruction without requiring fine-tuning, suggesting the potential of mathematically grounded dynamic-static disentanglement.

Lanyun Zhu, Deyi Ji, Tianrun Chen et al.

Despite achieving rapid developments and with widespread applications, Large Vision-Language Models (LVLMs) confront a serious challenge of being prone to generating hallucinations. An over-reliance on linguistic priors has been identified as a key factor leading to these hallucinations. In this paper, we propose to alleviate this problem by introducing a novel image-biased decoding (IBD) technique. Our method derives the next-token probability distribution by contrasting predictions from a conventional LVLM with those of an image-biased LVLM, thereby amplifying the correct information highly correlated with image content while mitigating the hallucinatory errors caused by excessive dependence on text. We further conduct a comprehensive statistical analysis to validate the reliability of our method, and design an adaptive adjustment strategy to achieve robust and flexible handling under varying conditions. Experimental results across multiple evaluation metrics verify that our method, despite not requiring additional training data and only with a minimal increase in model parameters, can significantly reduce hallucinations in LVLMs and enhance the truthfulness of the generated response.

Shangzan Zhang, Sida Peng, Tao Xu et al.

This paper aims to generate materials for 3D meshes from text descriptions. Unlike existing methods that synthesize texture maps, we propose to generate segment-wise procedural material graphs as the appearance representation, which supports high-quality rendering and provides substantial flexibility in editing. Instead of relying on extensive paired data, i.e., 3D meshes with material graphs and corresponding text descriptions, to train a material graph generative model, we propose to leverage the pre-trained 2D diffusion model as a bridge to connect the text and material graphs. Specifically, our approach decomposes a shape into a set of segments and designs a segment-controlled diffusion model to synthesize 2D images that are aligned with mesh parts. Based on generated images, we initialize parameters of material graphs and fine-tune them through the differentiable rendering module to produce materials in accordance with the textual description. Extensive experiments demonstrate the superior performance of our framework in photorealism, resolution, and editability over existing methods. Project page: https://zju3dv.github.io/MaPa

Lanyun Zhu, Tianrun Chen, Qianxiong Xu et al.

Existing LVLM-based reasoning segmentation methods often suffer from imprecise segmentation results and hallucinations in their text responses. This paper introduces POPEN, a novel framework designed to address these issues and achieve improved results. POPEN includes a preference-based optimization method to finetune the LVLM, aligning it more closely with human preferences and thereby generating better text responses and segmentation results. Additionally, POPEN introduces a preference-based ensemble method for inference, which integrates multiple outputs from the LVLM using a preference-score-based attention mechanism for refinement. To better adapt to the segmentation task, we incorporate several task-specific designs in our POPEN framework, including a new approach for collecting segmentation preference data with a curriculum learning mechanism, and a novel preference optimization loss to refine the segmentation capability of the LVLM. Experiments demonstrate that our method achieves state-of-the-art performance in reasoning segmentation, exhibiting minimal hallucination in text responses and the highest segmentation accuracy compared to previous advanced methods like LISA and PixelLM. Project page is https://lanyunzhu.site/POPEN/

Ying Zang, Chenglong Fu, Runlong Cao et al.

Referring expression segmentation (RES), a task that involves localizing specific instance-level objects based on free-form linguistic descriptions, has emerged as a crucial frontier in human-AI interaction. It demands an intricate understanding of both visual and textual contexts and often requires extensive training data. This paper introduces RESMatch, the first semi-supervised learning (SSL) approach for RES, aimed at reducing reliance on exhaustive data annotation. Extensive validation on multiple RES datasets demonstrates that RESMatch significantly outperforms baseline approaches, establishing a new state-of-the-art. Although existing SSL techniques are effective in image segmentation, we find that they fall short in RES. Facing the challenges including the comprehension of free-form linguistic descriptions and the variability in object attributes, RESMatch introduces a trifecta of adaptations: revised strong perturbation, text augmentation, and adjustments for pseudo-label quality and strong-weak supervision. This pioneering work lays the groundwork for future research in semi-supervised learning for referring expression segmentation.

Cheng Chen, Jiacheng Wei, Tianrun Chen et al.

Creating CAD digital twins from the physical world is crucial for manufacturing, design, and simulation. However, current methods typically rely on costly 3D scanning with labor-intensive post-processing. To provide a user-friendly design process, we explore the problem of reverse engineering from unconstrained real-world CAD images that can be easily captured by users of all experiences. However, the scarcity of real-world CAD data poses challenges in directly training such models. To tackle these challenges, we propose CADCrafter, an image-to-parametric CAD model generation framework that trains solely on synthetic textureless CAD data while testing on real-world images. To bridge the significant representation disparity between images and parametric CAD models, we introduce a geometry encoder to accurately capture diverse geometric features. Moreover, the texture-invariant properties of the geometric features can also facilitate the generalization to real-world scenarios. Since compiling CAD parameter sequences into explicit CAD models is a non-differentiable process, the network training inherently lacks explicit geometric supervision. To impose geometric validity constraints, we employ direct preference optimization (DPO) to fine-tune our model with the automatic code checker feedback on CAD sequence quality. Furthermore, we collected a real-world dataset, comprised of multi-view images and corresponding CAD command sequence pairs, to evaluate our method. Experimental results demonstrate that our approach can robustly handle real unconstrained CAD images, and even generalize to unseen general objects.

Lanyun Zhu, Tianrun Chen, Deyi Ji et al.

This paper proposes a new effective and efficient plug-and-play backbone for video-based person re-identification (ReID). Conventional video-based ReID methods typically use CNN or transformer backbones to extract deep features for every position in every sampled video frame. Here, we argue that this exhaustive feature extraction could be unnecessary, since we find that different frames in a ReID video often exhibit small differences and contain many similar regions due to the relatively slight movements of human beings. Inspired by this, a more selective, efficient paradigm is explored in this paper. Specifically, we introduce a patch selection mechanism to reduce computational cost by choosing only the crucial and non-repetitive patches for feature extraction. Additionally, we present a novel network structure that generates and utilizes pseudo frame global context to address the issue of incomplete views resulting from sparse inputs. By incorporating these new designs, our backbone can achieve both high performance and low computational cost. Extensive experiments on multiple datasets show that our approach reduces the computational cost by 74\% compared to ViT-B and 28\% compared to ResNet50, while the accuracy is on par with ViT-B and outperforms ResNet50 significantly.

Ying Zang, Yidong Han, Chaotao Ding et al.

Reconstructing dynamic 4D scenes is an important yet challenging task. While 3D foundation models like VGGT excel in static settings, they often struggle with dynamic sequences where motion causes significant geometric ambiguity. To address this, we present a framework designed to disentangle dynamic and static components by modeling uncertainty across different stages of the reconstruction process. Our approach introduces three synergistic mechanisms: (1) Entropy-Guided Subspace Projection, which leverages information-theoretic weighting to adaptively aggregate multi-head attention distributions, effectively isolating dynamic motion cues from semantic noise; (2) Local-Consistency Driven Geometry Purification, which enforces spatial continuity via radius-based neighborhood constraints to eliminate structural outliers; and (3) Uncertainty-Aware Cross-View Consistency, which formulates multi-view projection refinement as a heteroscedastic maximum likelihood estimation problem, utilizing depth confidence as a probabilistic weight. Experiments on dynamic benchmarks show that our approach outperforms current state-of-the-art methods, reducing Mean Accuracy error by 13.43\% and improving segmentation F-measure by 10.49\%. Our framework maintains the efficiency of feed-forward inference and requires no task-specific fine-tuning or per-scene optimization.

Deyi Ji, Yuekui Yang, Haiyang Wu et al.

Advertisement (Ad) video violation detection is critical for ensuring platform compliance, but existing methods struggle with precise temporal grounding, noisy annotations, and limited generalization. We propose RAVEN, a novel framework that integrates curriculum reinforcement learning with multimodal large language models (MLLMs) to enhance reasoning and cognitive capabilities for violation detection. RAVEN employs a progressive training strategy, combining precisely and coarsely annotated data, and leverages Group Relative Policy Optimization (GRPO) to develop emergent reasoning abilities without explicit reasoning annotations. Multiple hierarchical sophisticated reward mechanism ensures precise temporal grounding and consistent category prediction. Experiments on industrial datasets and public benchmarks show that RAVEN achieves superior performances in violation category accuracy and temporal interval localization. We also design a pipeline to deploy the RAVEN on the online Ad services, and online A/B testing further validates its practical applicability, with significant improvements in precision and recall. RAVEN also demonstrates strong generalization, mitigating the catastrophic forgetting issue associated with supervised fine-tuning.

Lanyun Zhu, Deyi Ji, Tianrun Chen et al.

The success of DeepSeek-R1 demonstrates the immense potential of using reinforcement learning (RL) to enhance LLMs' reasoning capabilities. This paper introduces Retrv-R1, the first R1-style MLLM specifically designed for multimodal universal retrieval, achieving higher performance by employing step-by-step reasoning to produce more accurate retrieval results. We find that directly applying the methods of DeepSeek-R1 to retrieval tasks is not feasible, mainly due to (1) the high computational cost caused by the large token consumption required for multiple candidates with reasoning processes, and (2) the instability and suboptimal results when directly applying RL to train for retrieval tasks. To address these issues, Retrv-R1 introduces an information compression module with a details inspection mechanism, which enhances computational efficiency by reducing the number of tokens while ensuring that critical information for challenging candidates is preserved. Furthermore, a new training paradigm is proposed, including an activation stage using a retrieval-tailored synthetic CoT dataset for more effective optimization, followed by RL with a novel curriculum reward to improve both performance and efficiency. Incorporating these novel designs, Retrv-R1 achieves SOTA performance, high efficiency, and strong generalization ability, as demonstrated by experiments across multiple benchmarks and tasks.

Ying Zang, Yuncan Gao, Jiangi Zhang et al.

This work advances zero-shot interactive segmentation for remote sensing imagery through three key contributions. First, we propose a novel sketch-based prompting method, enabling users to intuitively outline objects, surpassing traditional point or box prompts. Second, we introduce LTL-Sensing, the first dataset pairing human sketches with remote sensing imagery, setting a benchmark for future research. Third, we present LTL-Net, a model featuring a multi-input prompting transport module tailored for freehand sketches. Extensive experiments show our approach significantly improves segmentation accuracy and robustness over state-of-the-art methods like SAM, fostering more intuitive human-AI collaboration in remote sensing analysis and enhancing its applications.

Tianrun Chen, Runlong Cao, Xinda Yu et al.

The rapid rise of large-scale foundation models has reshaped the landscape of image segmentation, with models such as Segment Anything achieving unprecedented versatility across diverse vision tasks. However, previous generations-including SAM and its successor-still struggle with fine-grained, low-level segmentation challenges such as camouflaged object detection, medical image segmentation, cell image segmentation, and shadow detection. To address these limitations, we originally proposed SAM-Adapter in 2023, demonstrating substantial gains on these difficult scenarios. With the emergence of Segment Anything 3 (SAM3)-a more efficient and higher-performing evolution with a redesigned architecture and improved training pipeline-we revisit these long-standing challenges. In this work, we present SAM3-Adapter, the first adapter framework tailored for SAM3 that unlocks its full segmentation capability. SAM3-Adapter not only reduces computational overhead but also consistently surpasses both SAM and SAM2-based solutions, establishing new state-of-the-art results across multiple downstream tasks, including medical imaging, camouflaged (concealed) object segmentation, and shadow detection. Built upon the modular and composable design philosophy of the original SAM-Adapter, SAM3-Adapter provides stronger generalizability, richer task adaptability, and significantly improved segmentation precision. Extensive experiments confirm that integrating SAM3 with our adapter yields superior accuracy, robustness, and efficiency compared to all prior SAM-based adaptations. We hope SAM3-Adapter can serve as a foundation for future research and practical segmentation applications. Code, pre-trained models, and data processing pipelines are available.

Deyi Ji, Yuekui Yang, Liqun Liu et al.

Advertising (Ad) is a cornerstone of the digital economy, yet the moderation of video advertisements remains a significant challenge due to their complexity and the need for precise violation localization. While recent advancements, such as the RAVEN model, have improved coarse-grained violation detection, critical gaps persist in fine-grained understanding, explainability, and generalization. To address these limitations, we propose RAVEN++, a novel framework that introduces three key innovations: 1) Active Reinforcement Learning (RL), which dynamically adapts training to samples of varying difficulty; 2) Fine-Grained Violation Understanding, achieved through hierarchical reward functions and reasoning distillation; and 3) Progressive Multi-Stage Training, which systematically combines knowledge injection, curriculum-based passive RL, and active RL. Extensive experiments on both public and proprietary datasets, on both offline scenarios and online deployed A/B Testing, demonstrate that RAVEN++ outperforms general-purpose LLMs and specialized models like RAVEN in terms of fine-grained violation understanding, reasoning capabilities, and generalization ability.

Ying Zang, Yuanqi Hu, Xinyu Chen et al.

In the era of immersive consumer electronics, such as AR/VR headsets and smart devices, people increasingly seek ways to express their identity through virtual fashion. However, existing 3D garment design tools remain inaccessible to everyday users due to steep technical barriers and limited data. In this work, we introduce a 3D sketch-driven 3D garment generation framework that empowers ordinary users - even those without design experience - to create high-quality digital clothing through simple 3D sketches in AR/VR environments. By combining a conditional diffusion model, a sketch encoder trained in a shared latent space, and an adaptive curriculum learning strategy, our system interprets imprecise, free-hand input and produces realistic, personalized garments. To address the scarcity of training data, we also introduce KO3DClothes, a new dataset of paired 3D garments and user-created sketches. Extensive experiments and user studies confirm that our method significantly outperforms existing baselines in both fidelity and usability, demonstrating its promise for democratized fashion design on next-generation consumer platforms.

Chunan Yu, Yidong Han, Chaotao Ding et al.

In the era of the metaverse, where immersive technologies redefine human experiences, translating abstract literary concepts into navigable 3D environments presents a fundamental challenge in preserving semantic and emotional fidelity. This research introduces HaikuVerse, a novel framework for transforming poetic abstraction into spatial representation, with Japanese Haiku serving as an ideal test case due to its sophisticated encapsulation of profound emotions and imagery within minimal text. While existing text-to-3D methods struggle with nuanced interpretations, we present a literary-guided approach that synergizes traditional poetry analysis with advanced generative technologies. Our framework centers on two key innovations: (1) Hierarchical Literary-Criticism Theory Grounded Parsing (H-LCTGP), which captures both explicit imagery and implicit emotional resonance through structured semantic decomposition, and (2) Progressive Dimensional Synthesis (PDS), a multi-stage pipeline that systematically transforms poetic elements into coherent 3D scenes through sequential diffusion processes, geometric optimization, and real-time enhancement. Extensive experiments demonstrate that HaikuVerse significantly outperforms conventional text-to-3D approaches in both literary fidelity and visual quality, establishing a new paradigm for preserving cultural heritage in immersive digital spaces. Project website at: https://syllables-to-scenes.github.io/