论文检索

YuEun Lee, Jung Uk Kim

Video moment retrieval (MR) and highlight detection (HD) with natural language queries aim to localize relevant moments and key highlights in a video clips. However, existing methods overlook the importance of individual words, treating the entire text query and video clips as a black-box, which hinders contextual understanding. In this paper, we propose a novel approach that enables fine-grained clip filtering by identifying and prioritizing important words in the query. Our method integrates image-text scene understanding through Multimodal Large Language Models (MLLMs) and enhances the semantic understanding of video clips. We introduce a feature enhancement module (FEM) to capture important words from the query and a ranking-based filtering module (RFM) to iteratively refine video clips based on their relevance to these important words. Extensive experiments demonstrate that our approach significantly outperforms existing state-of-the-art methods, achieving superior performance in both MR and HD tasks.

Yongjoon Lee, Donghyeon Cho

Object removal in 3D space is a key technology for immersive applications such as virtual reality (VR), augmented reality (AR), and the metaverse. While recent approaches have attempted to address this task using 2D inpainting models, they often suffer from two major limitations: (1) inaccurate geometric restoration in the removed regions, and (2) visual inconsistency across multiple viewpoints. To address these challenges, we propose GPGS, a novel pipeline built upon the 3D Gaussian Splatting (3DGS) framework. First, we perform geometry-aware 3D inpainting by leveraging a pre-trained point cloud completion model and a coarse-to-fine inference strategy, enabling accurate restoration of unseen 3D structures. Next, we introduce a projected image refinement method that improves the appearance of novel-view projections by addressing view-dependent artifacts such as brightness shifts and texture misalignments. GPGS further enhances overall scene consistency through fine-tuning of the original 3DGS scene using the refined multi-view images. Experimental results show that our GPGS makes geometrically accurate and visually coherent outputs, even in challenging 360° panoramic scenes, significantly outperforming existing methods.

Soyul Lee, Seungmin Baek, Dongbo Min

Monocular 3D object detection is a cost-effective solution for applications like autonomous driving and robotics, but remains fundamentally ill-posed due to inherently ambiguous depth cues. Recent DETR-based methods attempt to mitigate this through global attention and auxiliary depth prediction, yet they still struggle with inaccurate depth estimates. Moreover, these methods often overlook instance-level detection difficulty, such as occlusion, distance, and truncation, leading to suboptimal detection performance. We propose MonoDLGD, a novel Difficulty-Aware Label-Guided Denoising framework that adaptively perturbs and reconstructs ground-truth labels based on detection uncertainty. Specifically, MonoDLGD applies stronger perturbations to easier instances and weaker ones into harder cases, and then reconstructs them to effectively provide explicit geometric supervision. By jointly optimizing label reconstruction and 3D object detection, MonoDLGD encourages geometry-aware representation learning and improves robustness to varying levels of object complexity. Extensive experiments on the KITTI benchmark demonstrate that MonoDLGD achieves state-of-the-art performance across all difficulty levels.

Minsung Lee, Seo Hyun Kim, Yeonsu Park, Hyeongseok Seo, Jongmin Lee

Accurate and efficient depth estimation from time-of-flight (ToF) LiDAR is essential for autonomous systems operating in real-world environments. However, traditional histogram-based depth estimation (HBDE) algorithms face fundamental limitations in balancing depth performance and computational cost, and they struggle under signal-induced pile-up distortion. While deep learning has shown promise, existing neural network-based methods rely on large models that are impractical for deployment on edge hardware. To bridge this critical gap, we propose a paradigm shift in histogram-based ToF estimation, reframing depth estimation from signal filtering to lightweight similarity learning. Instead of attempting to correct the distorted signal, our approach learns a specialized metric where the measure of similarity between the distorted histogram and a reference pulse is the temporal shift itself. The resulting 57.61 KB model, over 215.2 times smaller than state-of-the-art deep learning approaches, achieves real-time performance (106.27 fps) on an FPGA. It delivers superior accuracy across nearly all signal-noise conditions, including 2.21 cm RMSE at severe pile-up scenarios, significantly outperforming conventional methods while remaining practical for on-device deployment.

Junhee Lee, ChaeBeen Bang, MyoungChul Kim, MyeongAh Cho

Weakly-Supervised Video Anomaly Detection aims to identify anomalous events using only video-level labels, balancing annotation efficiency with practical applicability. However, existing methods often oversimplify the anomaly space by treating all abnormal events as a single category, overlooking the diverse semantic and temporal characteristics intrinsic to real-world anomalies. Inspired by how humans perceive anomalies, by jointly interpreting temporal motion patterns and semantic structures underlying different anomaly types, we propose RefineVAD, a novel framework that mimics this dual-process reasoning. Our framework integrates two core modules. The first, Motion-aware Temporal Attention and Recalibration (MoTAR), estimates motion salience and dynamically adjusts temporal focus via shift-based attention and global Transformer-based modeling. The second, Category-Oriented Refinement (CORE), injects soft anomaly category priors into the representation space by aligning segment-level features with learnable category prototypes through cross-attention. By jointly leveraging temporal dynamics and semantic structure, explicitly models both ``how'' motion evolves and ``what'' semantic category it resembles. Extensive experiments on WVAD benchmark validate the effectiveness of RefineVAD and highlight the importance of integrating semantic context to guide feature refinement toward anomaly-relevant patterns.

JoungBin Lee, Hyunkoo Lee, Jini Yang, Chaehyun Kim, Jung Yi, Seok Hwangbo, Hyeoncheol Lee, Minho Chun, Eunjo Jeong, Seungryong Kim

We present CHIMERA, a novel framework for generating realistic, generalizable, and prompt-driven industrial anomalies from natural language instructions. Our method addresses two key challenges in text-guided anomaly synthesis: (1) the scarcity of scalable, high-quality paired anomaly data and (2) the difficulty of efficiently adapting large diffusion models to domain-specific tasks without overfitting. To tackle these challenges, we first introduce a Vision-Language Model (VLM)-guided data curation pipeline that automatically generates semantically rich and spatially grounded captions from normal images, enabling effective dataset augmentation without manual annotations. Building upon this, we propose a parameter-efficient fine-tuning strategy that adapts a pre-trained Diffusion Transformer (Stable Diffusion 3) using lightweight LoRA adapters. By aligning structured prompts with the model's pre-trained language-vision prior and introducing auxiliary attention-based mask supervision, our method prevents overfitting, enhances spatial consistency, and ensures efficient training even with limited data. Extensive experiments show that CHIMERA is the first unified framework to achieve controllable, scalable, and generalizable industrial anomaly generation by integrating VLM-guided data curation with efficient diffusion-based training, significantly improving anomaly detection in low-data and unseen scenarios.

Jeonghaeng Lee, Seok Keun Choi, Zhixuan Li, Weisi Lin, Sanghoon Lee

While recent 3D head avatar creation methods attempt to animate facial dynamics, they often fail to capture personalized details, limiting realism and expressiveness. To fill this gap, we present DipGuava (Disentangled and Personalized Gaussian UV Avatar), a novel 3D Gaussian head avatar creation method that successfully generates avatars with personalized attributes from monocular video. DipGuava is the first method to explicitly disentangle facial appearance into two complementary components, trained in a structured two-stage pipeline that significantly reduces learning ambiguity and enhances reconstruction fidelity. In the first stage, we learn a stable geometry-driven base appearance that captures global facial structure and coarse expression-dependent variations. In the second stage, the personalized residual details not captured in the first stage are predicted, including high-frequency components and nonlinearly varying features such as wrinkles and subtle skin deformations. These components are fused via dynamic appearance fusion that integrates residual details after deformation, ensuring spatial and semantic alignment. This disentangled design enables DipGuava to generate photorealistic, identity-preserving avatars, consistently outperforming prior methods in both visual quality and quantitative performance, as demonstrated in extensive experiments.

Jae Joong Lee, Bedrich Benes, Raymond A. Yeh

Amodal segmentation is an image-based algorithm that aims to predict masks for both visible and occluded parts of objects. Existing methods typically rely on supervised learning with annotated amodal masks or synthetic data. The effectiveness of these methods relies heavily on the quality of the datasets. This dependence can unintentionally restrict their generalization capabilities due to insufficient diversity and size. Although existing zero-shot methods perform well on their reported datasets, their performance does not necessarily transfer to other datasets. We propose a tuning-free approach that re-purposes diffusion-based inpainting foundation models for amodal segmentation. Our approach is motivated by the “occlusion-free bias” of inpainting models, i.e., the inpainted objects tend to be complete and without occlusions. We reconstruct the occluded regions of an object via inpainting and then apply segmentation, all without additional training or fine-tuning. Experiments on five datasets, three previously unreported, demonstrate the generalizability of our approach. On average, our approach achieves 5.3% more accurate masks in mIoU compared to the publicly available state-of-the-art, pix2gestalt.

Hyundo Lee, Suhyung Choi, Inwoo Hwang, Byoung-Tak Zhang

Image generation models trained on large datasets can synthesize high-quality images but often produce spatially inconsistent and distorted images due to limited information about the underlying structures and spatial layouts. In this work, we leverage intrinsic scene properties (e.g., depth, segmentation maps) that provide rich information about the underlying scene, unlike prior approaches that solely rely on image-text pairs or use intrinsics as conditional inputs. Our approach aims to co-generate both images and their corresponding intrinsics, enabling the model to implicitly capture the underlying scene structure and generate more spatially consistent and realistic images. Specifically, we first extract rich intrinsic scene properties from a large image dataset with pre-trained estimators, eliminating the need for additional scene information or explicit 3D representations. We then aggregate various intrinsic scene properties into a single latent variable using an autoencoder. Building upon pre-trained large-scale Latent Diffusion Models (LDMs), our method simultaneously denoises the image and intrinsic domains by carefully sharing mutual information so that the image and intrinsic reflect each other without degrading image quality. Experimental results demonstrate that our method corrects spatial inconsistencies and produces a more natural layout of scenes while maintaining the fidelity and textual alignment of the base model (e.g., Stable Diffusion).

Hojun Lee, Mijin Koo, Yeji Song, Nojun Kwak

Recent advancements in diffusion models have made fine-tuning text-to-image models for personalization increasingly accessible, but have also raised significant concerns regarding unauthorized data usage and privacy infringement. Current protection methods are limited to passively degrading image quality, failing to achieve stable control. While Targeted Data Protection (TDP) offers a promising paradigm for active redirection toward user-specified target concepts, existing TDP attempts suffer from poor controllability due to snapshot-matching approaches that fail to account for complete learning dynamics. We introduce TAFAP (Trajectory Alignment via Fine-tuning with Adversarial Perturbations), the first method to successfully achieve effective TDP by controlling the entire training trajectory. Unlike snapshot-based methods whose protective influence is easily diluted as training progresses, TAFAP employs trajectory-matching inspired by dataset distillation to enforce persistent, verifiable transformations throughout fine-tuning. We validate our method through extensive experiments, demonstrating the first successful targeted transformation in diffusion models with simultaneous control over both identity and visual patterns. TAFAP significantly outperforms existing TDP attempts, achieving robust redirection toward target concepts while maintaining high image quality. This work enables verifiable safeguards and provides a new framework for controlling and tracing alterations in diffusion model outputs.

Jiajun Le, Jiayi Ma

Recent progress in two-view geometry increasingly emphasizes enforcing smoothness and global consistency priors when estimating motion fields between pairs of images. However, in complex real-world scenes, characterized by extreme viewpoint and scale changes as well as pronounced depth discontinuities, the motion field often exhibits diverse and heterogeneous motion patterns. Most existing methods lack targeted modeling strategies and fail to explicitly account for this variability, resulting in estimated motion fields that diverge from their true underlying structure and distribution. We observe that Mixture-of-Experts (MoE) can assign dedicated experts to motion sub-fields, enabling a divide-and-conquer strategy for heterogeneous motion patterns. Building on this insight, we re-architect motion field modeling in two-view geometry with GeoMoE, a streamlined framework. Specifically, we first devise a Probabilistic Prior-Guided Decomposition strategy that exploits inlier probability signals to perform a structure-aware decomposition of the motion field into heterogeneous sub-fields, sharply curbing outlier-induced bias. Next, we introduce an MoE-Enhanced Bi-Path Rectifier that enhances each sub-field along spatial-context and channel-semantic paths and routes it to a customized expert for targeted modeling, thereby decoupling heterogeneous motion regimes, suppressing cross-sub-field interference and representational entanglement, and yielding fine-grained motion-field rectification. With this minimalist design, GeoMoE outperforms prior state-of-the-art methods in relative pose and homography estimation and shows strong generalization.

Duy-Tho Le, Trung Pham, Jianfei Cai, Hamid Rezatofighi

Optimizing the similarity between parametric shapes is crucial for numerous computer vision tasks, where Intersection over Union (IoU) stands as the canonical measure. However, existing optimization methods exhibit significant shortcomings: regression-based losses like L1/L2 lack correlation with IoU, IoU-based losses are unstable and limited to simple shapes, and task-specific methods are computationally intensive and not generalizable across domains. As a result, the current landscape of parametric shape objective functions has become scattered, with each domain proposing distinct IoU approximations. To address this, we unify the parametric shape optimization objective functions by introducing Marginalized Generalized IoU (MGIoU), a novel loss function that overcomes these challenges by projecting structured convex shapes onto their unique shape Normals to compute one-dimensional normalized GIoU. MGIoU offers a simple, efficient, fully differentiable approximation strongly correlated with IoU. We extend MGIoU to MGIoU+ that supports optimizing unstructured convex shapes. Together, MGIoU and MGIoU+ unify parametric shape optimization across diverse applications. Experiments on standard benchmarks demonstrate that MGIoU and MGIoU+ demonstrate higher performance while reducing loss computation latency up to 10-40x. Also, MGIoU and MGIoU+ satisfy metric properties and scale-invariance, ensuring robustness as an objective function. We further propose MGIoU- for minimizing overlaps in tasks like collision-free trajectory prediction.

Maria Larchenko, Dmitry Guskov, Alexander Lobashev, Georgy Derevyanko

Color harmonization adjusts the colors of an inserted object so that it perceptually matches the surrounding image, resulting in a seamless composite. The harmonization problem naturally arises in augmented reality (AR), yet harmonization algorithms are not currently integrated into AR pipelines because real-time solutions are scarce. In this work, we address color harmonization for AR by proposing a lightweight approach that supports on-device inference. For this, we leverage classical optimal transport theory by training a compact encoder to predict the Monge-Kantorovich transport map. We benchmark our MKL-Harmonizer algorithm against state-of-the-art methods and demonstrate that for real composite AR images our method achieves the best aggregated score. We release our dedicated AR dataset of composite images with pixel-accurate masks and data-gathering toolkit to support further data acquisition by researchers.

Tianzhong Lan, Zhang Yi, Xiuyuan Xu, Min Zhu

Data economics drives AI by optimizing data usage, reducing costs, and enhancing efficiency. In 3D tumor segmentation, efficiency is crucial due to the high demand for labor-intensive manual annotations. Box-supervised segmentation offers a promising alternative but is constrained by tumor morphology complexity and boundary ambiguity. In this paper, we propose a novel 3D tumor segmentation model that integrates both positional and embedding features to facilitate inter-task collaboration. We introduce an Anatomical-Driven Class Activation Map to predefine the complex tumor morphology prior, which is further refined by our Geometric Pixel Co-embedding Learner. This learner utilizes contrastive learning to encode semantic information between center and edge pixels, enhancing pixel clustering and progressively refining tumor boundary segmentation in a coarse-to-fine manner. Our approach outperforms existing box-supervised methods in segmentation performance, with extensive experiments on four tumor datasets demonstrating significant improvements. This work provides a cost-effective and efficient solution for tumor segmentation, advancing the application of data economics in medical imaging.

Meng Lan, Lefei Zhang, Xiaomeng Li

The Segment Anything Model 2 (SAM2) has demonstrated remarkable promptable visual segmentation capabilities in video data, showing potential for extension to medical image segmentation (MIS) tasks involving 3D volumes and temporally correlated 2D image sequences. However, adapting SAM2 to MIS presents several challenges, including the need for extensive annotated medical data for fine-tuning and high-quality manual prompts, which are both labor-intensive and require intervention from medical experts. To address these challenges, we introduce OFL-SAM2, a prompt-free SAM2 framework for label-efficient MIS. Our core idea is to leverage limited annotated samples to train a lightweight mapping network that captures medical knowledge and transforms generic image features into target features, thereby providing additional discriminative target representations for each frame and eliminating the need for manual prompts. Crucially, the mapping network supports online parameter update during inference, enhancing the model’s generalization across test sequences. Technically, we introduce two key components: (1) an online few-shot learner that trains the mapping network to generate target features using limited data, and (2) an adaptive fusion module that dynamically integrates the target features with the memory-attention features generated by frozen SAM2, leading to accurate and robust target representation. Extensive experiments on three diverse MIS datasets demonstrate that OFL-SAM2 achieves state-of-the-art performance with limited training data.

Zhangyu Lai, Yilin Lu, Xinyang Li, Jianghang Lin, Yansong Qu, Ming Li, Liujuan Cao

Visual anomaly detection is limited by the lack of sufficient anomaly data. While existing anomaly synthesis methods have made remarkable progress, achieving both realism and diversity in synthesis remains a major obstacle. To address this, we propose AnomalyPainter, a novel framework that breaks the diversity-realism trade-off dilemma through synergizing Vision Language Large Model (VLLM), Latent Diffusion Model (LDM), and our newly introduced texture library Tex-9K. Tex-9K is a professional texture library containing 75 categories and 8792 texture assets crafted for diverse anomaly synthesis. Leveraging VLLM's general knowledge, reasonable anomaly text descriptions are generated for each industrial object and matched with relevant diverse textures from Tex-9K. These textures then guide the LDM via ControlNet to paint on normal images. Furthermore, we introduce Texture-Aware Latent Init to stabilize the natural-image-trained ControlNet for industrial images. Extensive experiments show that AnomalyPainter outperforms existing methods in realism, diversity, and generalization, achieving superior downstream performance.

Zengyuan Lai, Jiarui Yang, Songpengcheng Xia, Lizhou Lin, Lan Sun, Renwen Wang, Jianran Liu, Qi Wu, Ling Pei

Millimeter-wave radar offers a privacy-preserving and environment-robust alternative to vision-based sensing, enabling human motion analysis in challenging conditions such as low light, occlusions, rain, or smoke. However, its sparse point clouds pose significant challenges for semantic understanding. We present RadarLLM, the first framework that leverages large language models (LLMs) for human motion understanding from radar signals. RadarLLM introduces two key innovations: (1) a motion-guided radar tokenizer based on our Aggregate VQ-VAE architecture, integrating deformable body templates and masked trajectory modeling to convert spatial-temporal radar sequences into compact semantic tokens; and (2) a radar-aware language model that establishes cross-modal alignment between radar and text in a shared embedding space. To overcome the scarcity of paired radar-text data, we generate a realistic radar-text dataset from motion-text datasets with a physics-aware synthesis pipeline. Extensive experiments on both synthetic and real-world benchmarks show that RadarLLM achieves state-of-the-art performance, enabling robust and interpretable motion understanding under privacy and visibility constraints, even in adverse environments.

JuneHyoung Kwon, MiHyeon Kim, Eunju Lee, Yoonji Lee, Seunghoon Lee, YoungBin Kim

Machine unlearning, which enables a model to forget specific data, is crucial for ensuring data privacy and model reliability. However, its effectiveness can be severely undermined in real-world scenarios where models learn unintended biases from spurious correlations within the data. This paper investigates the unique challenges of unlearning from such biased models. We identify a novel phenomenon we term "shortcut unlearning," where models exhibit an "easy to learn, yet hard to forget" tendency. Specifically, models struggle to forget easily-learned, bias-aligned samples; instead of forgetting the class attribute, they unlearn the bias attribute, which can paradoxically improve accuracy on the class intended to be forgotten. To address this, we propose CUPID, a new unlearning framework inspired by the observation that samples with different biases exhibit distinct loss landscape sharpness. Our method first partitions the forget set into causal- and bias-approximated subsets based on sample sharpness, then disentangles model parameters into causal and bias pathways, and finally performs a targeted update by routing refined causal and bias gradients to their respective pathways. Extensive experiments on biased datasets including Waterbirds, BAR, and Biased NICO++ demonstrate that our method achieves state-of-the-art forgetting performance and effectively mitigates the shortcut unlearning problem.

Minseong Kweon, Jinsun Park

We introduce OceanSplat, a novel 3D Gaussian Splatting-based approach for high-fidelity underwater scene reconstruction. To overcome multi-view inconsistencies caused by scattering media, we design a trinocular setup for each camera pose by rendering from horizontally and vertically translated virtual viewpoints, enforcing view consistency to facilitate spatial optimization of 3D Gaussians. Furthermore, we derive synthetic epipolar depth priors from the virtual viewpoints, which serve as self-supervised depth regularizers to compensate for the limited geometric cues in degraded underwater scenes. We also propose a depth-aware alpha adjustment that modulates the opacity of 3D Gaussians during early training based on their depth along the viewing direction, deterring the formation of medium-induced primitives. Our approach promotes the disentanglement of 3D Gaussians from the scattering medium through effective geometric constraints, enabling accurate representation of scene structure and significantly reducing floating artifacts. Experiments on real-world underwater and simulated scenes demonstrate that OceanSplat substantially outperforms existing methods for both scene reconstruction and restoration in scattering media.

Yogesh Kumar, Anand Mishra

Few-shot Video Object Detection (FSVOD) addresses the challenge of detecting novel objects in videos with limited labeled examples, overcoming the constraints of traditional detection methods that require extensive training data. This task presents key challenges, including maintaining temporal consistency across frames affected by occlusion and appearance variations, and achieving novel object generalization without relying on complex region proposals, which are often computationally expensive and require task-specific training. Our novel object-aware temporal modeling approach addresses these challenges by incorporating a filtering mechanism that selectively propagates high-confidence object features across frames. This enables efficient feature progression, reduces noise accumulation, and enhances detection accuracy in a few-shot setting. By utilizing few-shot trained detection and classification heads with focused feature propagation, we achieve robust temporal consistency without depending on explicit object tube proposals. Our approach achieves performance gains, with AP improvements of 3.7% (FSVOD-500), 5.3% (FSYTV-40), 4.3% (VidOR), and 4.5% (VidVRD) in the 5-shot setting. Further results demonstrate improvements in 1-shot, 3-shot, and 10-shot configurations.