WACV 2026 论文列表

Lesion segmentation is an essential task in medical imaging to support diagnosis and assessment of pathologies. While deep learning models have shown success in various domains, their reliance on large-scale annotated datasets limits applicability in the medical domain due to labeling cost. To address this issue, recent studies in medical image segmentation have utilized clinical texts as complementary semantic cues without additional annotations. However, most existing methods utilize a single textual embedding and fail to capture hierarchical interactions between language and visual features, which limits their ability to leverage finegrained cues essential for precise and detailed segmentation. In this regime, we propose Hierarchical Visual-Textual Mixing Network (HiMix), a novel multi-modal segmentation framework that mixes multi-scale image and text representations throughout the mask decoding process. HiMix progressively injects hierarchical text embedding, from highlevel semantics to fine-grained spatial details, into corresponding image decoder layers to bridge the modality gap and enhance visual feature refinement at multiple levels of abstraction. Experiments on the QaTa-COV19, MosMed- Data+ and Kvasir-SEG datasets demonstrate that HiMix consistently outperforms uni-modal and multi-modal methods. Furthermore, HiMix exhibits strong generalization to unstructured textual formats, highlighting its practical applicability in real-world clinical scenarios.

Transparent image layer generation plays a significant role in digital art and design workflows. Existing methods typically decompose transparent layers from a single RGB image using a set of tools or generate multiple transparent layers sequentially. Despite some promising results, these methods often limit their ability to model global layout, physically plausible interactions, and visual effects such as shadows and reflections with high alpha quality due to limited shared global context among layers. To address this issue, we propose PSDiffusion, a unified diffusion framework that leverages image composition priors from pre-trained image diffusion model for simultaneous multi-layer text-to-image generation. Specifically, our method introduces a global layer interaction mechanism to generate layered images collaboratively, ensuring both individual layer quality and coherent spatial and visual relationships across layers. We include extensive experiments on benchmark datasets to demonstrate that PSDiffusion is able to outperform existing methods in generating multi-layer images with plausible structure and enhanced visual fidelity.

Evaluating the perceptual quality of Novel View Synthesis (NVS) images remains a key challenge, particularly in the absence of pixel-aligned ground truth references. Full-Reference Image Quality Assessment (FR-IQA) methods fail under misalignment, while No-Reference (NR-IQA) methods struggle with generalization. In this work, we introduce a Non-Aligned Reference (NAR-IQA) framework tailored for NVS, where it is assumed that the reference view shares partial scene content but lacks pixel-level alignment. We constructed a large-scale image dataset containing synthetic distortions targeting Temporal Regions of Interest (TROI) to train our NAR-IQA model. Our model is built on a contrastive learning framework that incorporates LoRA-enhanced DINOv2 embeddings and is guided by supervision from existing IQA methods. We train exclusively on synthetically generated distortions, deliberately avoiding overfitting to specific real NVS samples and thereby enhancing the model's generalization capability. Our model outperforms state-of-the-art FR-IQA, NR-IQA, and NAR-IQA methods, achieving robust performance on both aligned and non-aligned references. We also conducted a novel user study to gather data on human preferences when viewing non-aligned references in NVS. We find strong correlation between our proposed quality prediction model and the collected subjective ratings.

Although Multimodal Large Language Models (MLLMs) have advanced substantially, they remain vulnerable to object hallucination caused by language priors and visual information loss. To address this, we propose SAVE (Sparse Autoencoder-Driven Visual Information Enhancement), a framework that mitigates hallucination by steering the model along Sparse Autoencoder (SAE) latent features. A binary object-presence question-answering probe identifies the SAE features most indicative of the model's visual information processing, referred to as visual understanding features. Steering the model along these identified features reinforces grounded visual understanding and effectively reduces hallucination. With its simple design, SAVE outperforms state-of-the-art training-free methods on standard benchmarks, achieving a 10%p improvement in CHAIR_S and consistent gains on POPE and MMHal-Bench. Extensive evaluations across multiple models and layers confirm the robustness and generalizability of our approach. Further analysis reveals that steering along visual understanding features suppresses the generation of uncertain object tokens and increases attention to image tokens, mitigating hallucination. Code is released at https://github.com/wiarae/SAVE.

Accurate thyroid nodule segmentation in ultrasound images is essential for effective diagnosis and treatment planning. While multitask learning has shown promise in improving segmentation performance, several challenges remain unresolved: (a) scarcity of labeled data, (b) lack of integration of domain-specific prior knowledge, and (c) limited robustness in real-world clinical scenarios. To address these issues, we propose SSMT-Net, a Semi-Supervised Multi-task Transformer-based Network that leverages unlabeled data for an initial unsupervised pre-training phase. In the subsequent supervised phase, our model jointly optimizes thyroid nodule segmentation, thyroid gland segmentation, and nodule size estimation, effectively integrating both local and global contextual cues. This multitask formulation enables the model to generalize better and remain robust across variable clinical conditions. Evaluated on two public datasets, TN3K and DDTI, SSMT-Net sets a new benchmark in thyroid nodule segmentation, achieving up to 3.38% and 1.23% absolute improvements in IoU and DSC, respectively, compared to existing state-of-the-art methods. Our code is available at: https://github.com/Umar-Faroq/SSMT-Net.

Generating realistic human motions that naturally respond to both spoken language and physical objects is crucial for interactive digital experiences. Current methods, however, address speech-driven gestures or object interactions independently, limiting real-world applicability due to a lack of integrated, comprehensive datasets. To overcome this, we introduce InteracTalker, a novel framework that seamlessly integrates prompt-based object-aware interactions with co-speech gesture generation. We achieve this by employing a multi-stage training process to learn a unified motion, speech, and prompt embedding space. To support this, we curate a rich human-object interaction dataset, formed by augmenting an existing text-to-motion dataset with detailed object interaction annotations. Our framework utilizes a Generalized Motion Adaptation Module that enables independent training, adapting to the corresponding motion condition, which is then dynamically combined during inference. To address the imbalance between heterogeneous conditioning signals, we propose an adaptive fusion strategy, which dynamically reweights the conditioning signals during diffusion sampling.InteracTalker successfully unifies these previously separate tasks, outperforming prior methods in both co-speech gesture generation and object-interaction synthesis, outperforming gesture-focused diffusion methods, yielding highly realistic, object-aware full-body motions with enhanced realism, flexibility, and control.

Accurate geometric reconstruction of deformable tissues in monocular endoscopic video remains a fundamental challenge in robot-assisted minimally invasive surgery. Although recent volumetric and point primitive methods based on neural radiance fields (NeRF) and 3D Gaussian primitives have efficiently rendered surgical scenes, they still struggle with handling artifact-free tool occlusions and preserving fine anatomical details. These limitations stem from unrestricted Gaussian scaling and insufficient surface alignment constraints during reconstruction. To address these issues, we introduce Surgical Gaussian Surfels (SGS), which transform anisotropic point primitives into surface-aligned elliptical splats by constraining the scale component of the Gaussian covariance matrix along the view-aligned axis. We also introduce the Fully Fused Deformation Multilayer Perceptron (FFD-MLP), a lightweight Multi-Layer Perceptron (MLP) that predicts accurate surfel motion fields up to 5x faster than a standard MLP. This is coupled with locality constraints to handle complex tissue deformations. We use homodirectional view-space positional gradients to capture fine image details by splitting Gaussian Surfels in over-reconstructed regions. In addition, we define surface normals as the direction of the steepest density change within each Gaussian surfel primitive, enabling accurate normal estimation without requiring monocular normal priors. We evaluate our method on two in-vivo surgical datasets, where it outperforms current state-of-the-art methods in surface geometry, normal map quality, and rendering efficiency, while remaining competitive in real-time rendering performance. Code is available at https://git.new/NexyiHu.

We present CURIO, an OCR system for low-resource historical manuscripts. In many challenging cases, manuscripts feature curved text lines, unsegmented lines with lack of spacing between words, and highly variable line lengths -- conditions under which existing OCR methods fail. To tackle this challenge, we first extract lines and corresponding curvature profiles from manuscripts, then straighten them using a rectification procedure to reduce redundant background within each line. Because data is scarce, we compliment real data with synthetic data. To bridge the synthetic-real gap, we generate line images by warping rendered straight text along the rectified profiles, ensuring both real and synthetic lines align in their curvature characteristics. Our recognizer is a lightweight CNN-Transformer with padding-aware null activations, sparse attention and optimized with CTC loss for efficient training. We evaluate our method on challenging manuscript collections written in Sharada, a rare and endangered Indic script. CURIO outperforms strong CNN+RNN and Transformer baselines, with the largest gains on high-curvature lines and long lines. CURIO further transfers zero-shot to printed Sharada text, indicating robustness beyond manuscript domain.

With recent advancements in video backbone architectures and the remarkable success of large language models (LLMs), long-form video understanding--analyzing videos that span tens of minutes--has become both feasible and increasingly popular. However, the inherently redundant nature of video sequences presents significant challenges for current state-of-the-art models. These challenges arise from two key aspects: 1) efficiently incorporating a larger number of frames within the memory budget, and 2) extracting discriminative information from the vast volume of input data. In this paper, we present a novel, end-to-end schema for long-form video understanding, featuring an information-density-based adaptive video sampler (AVS) and an autoencoder based spatiotemporal video compressor (SVC) integrated with a multimodal large language model (MLLM). Our proposed system offers two significant advantages: it adaptively and effectively captures essential information from video sequences with various duration, and it achieves high compression rates while preserving crucial discriminative information. The proposed framework achieves promising performance across a range of benchmarks, excelling in both long-form video understanding tasks and standard video understanding benchmarks. These results demonstrate the versatility and effectiveness of our approach, particularly in handling the complexities of the long video sequences.

Work zones are essential to maintain, repair and upgrade our roadways. However, they introduce complex, dynamic and challenging environments for autonomous vehicles to navigate safely. To help address this challenge, we introduce the first publicly available, large-scale, multimodal work zone dataset collected with an autonomous vehicle consisting of multiple synchronized lidars and high-resolution cameras. The dataset covers various work zone elements like cones, barrels and other channelizers. Our dataset, referred to as WorkZone3D, consists of 3D annotation boxes for these objects. We also propose a detailed auto-annotation pipeline that can produce high-quality 3D labels, even for rare classes which do not have pre-trained 3D object detection models to start with. We evaluate a camera+lidar-based deep learning model on this dataset, highlighting the critical role of sensor fusion in accurate 3D localization of such small objects at a distance, often having very few lidar points on them. Our results demonstrate the usefulness of our dataset for generalization to real-world scenarios. We will release both our code and data.

Accurate reconstruction and relighting of glossy objects remains a longstanding challenge, as object shape, material properties, and illumination are inherently difficult to disentangle. Existing neural rendering approaches often rely on simplified BRDF models or parameterizations that couple diffuse and specular components, which restrict faithful material recovery and limit relighting fidelity. We propose a relightable framework that integrates a microfacet BRDF with the specular-glossiness parameterization into 2D Gaussian Splatting with deferred shading. This formulation enables more physically consistent material decomposition, while diffusion-based priors for surface normals and diffuse color guide early-stage optimization and mitigate ambiguity. A coarse-to-fine environment map optimization accelerates convergence, and negative-only environment map clipping preserves high-dynamic-range specular reflections. Extensive experiments on complex, glossy scenes demonstrate that our method achieves high-quality geometry and material reconstruction, delivering substantially more realistic and consistent relighting under novel illumination compared to existing Gaussian splatting methods.

Image segmentation foundation models (SFMs) like Segment Anything Model (SAM) have achieved impressive zero-shot and interactive segmentation across diverse domains. However, they struggle to segment objects with certain structures, particularly those with dense, tree-like morphology and low textural contrast from their surroundings. These failure modes are crucial for understanding the limitations of SFMs in real-world applications. To systematically study this issue, we introduce interpretable metrics quantifying object tree-likeness and textural separability. On carefully controlled synthetic experiments and real-world datasets, we show that SFM performance (e.g., SAM, SAM 2, HQ-SAM) noticeably correlates with these factors. We attribute these failures to SFMs misinterpreting local structure as global texture, resulting in over-segmentation or difficulty distinguishing objects from similar backgrounds. Notably, targeted fine-tuning fails to resolve this issue, indicating a fundamental limitation. Our study provides the first quantitative framework for modeling the behavior of SFMs on challenging structures, offering interpretable insights into their segmentation capabilities.

Lifelong learning on Whole Slide Images (WSIs) aims to train or fine-tune a unified model sequentially on cancer-related tasks, reducing the resources and effort required for data transfer and processing, especially given the gigabyte-scale size of WSIs. In this paper, we introduce MergeSlide, a simple yet effective framework that treats lifelong learning as a model-merging problem by leveraging a vision-language pathology foundation model. When a new task arrives, it is 1) defined with class-aware prompts, 2) fine-tuned for a few epochs in a classifier-free manner, and 3) merged into a unified model using an orthogonal continual-merging strategy that preserves performance and mitigates catastrophic forgetting. For inference under the class-incremental learning (CLASS-IL) setting, where task identity is unknown, we introduce Task-to-Class Prompt-aligned (TCP) inference. Specifically, TCP first identifies the most relevant task using task-level prompts and then applies the corresponding class-aware prompts to generate predictions. To evaluate MergeSlide, we conduct experiments on a stream of six TCGA datasets. The results show that MergeSlide outperforms both rehearsal-based continual learning and vision-language zero-shot baselines. Code and data are available at https://github.com/caodoanh2001/MergeSlide.

Image restoration is essential for enhancing degraded images across computer vision tasks. However, most existing methods address only a single type of degradation (e.g., blur, noise, or haze) at a time, limiting their real-world applicability where multiple degradations often occur simultaneously. In this paper, we propose UniCoRN, a unified image restoration approach capable of handling multiple degradation types simultaneously using a multi-head diffusion model. Specifically, we uncover the potential of low-level visual cues extracted from images in guiding a controllable diffusion model for real-world image restoration, and design a multi-head control network adaptable via a mixture-of-experts strategy. We train our model without any prior assumption of specific degradations, through a smartly designed curriculum learning recipe. Additionally, we also introduce MetaRestore, a metalens imaging benchmark containing images with multiple degradations and artifacts. Extensive evaluations on several challenging datasets, including our benchmark, demonstrate that our method achieves significant performance gains and can robustly restore images with severe degradations.

Recent point cloud frame interpolation methods predict an interpolated frame through the merging of two intermediate frames constructed by scene flow estimation. However, these methods introduce errors due to generation since they adopt a generative approach to merge the frames, degrading frame interpolation performance. In this paper, we propose a point cloud frame interpolation method with time-aware point cloud sampling and a self-supervised learning strategy, termed TS-PCI. The proposed method introduces a time-aware learning-based point cloud sampling model to merge the two frames into a single frame in a non-generative approach. The proposed method also introduces an attention-based geometry refinement model to improve the geometric quality of the sampled point clouds. Furthermore, the proposed method adopts a self-supervised strategy that dynamically creates ground truth labels for point cloud sampling, allowing the models to be trained in an end-to-end manner. Experimental results on three large-scale datasets show that the proposed method achieves superior performance compared to state-of-the-art methods.

We present Layout Anything, a transformer-based framework for indoor layout estimation that adapts the OneFormer's universal segmentation architecture to geometric structure prediction. Our approach integrates OneFormer's task-conditioned queries and contrastive learning with two key modules: (1) a layout degeneration strategy that augments training data while preserving Manhattan-world constraints through topology-aware transformations, and (2) differentiable geometric losses that directly enforce planar consistency and sharp boundary predictions during training. By unifying these components in an end-to-end framework, the model eliminates complex post-processing pipelines while achieving high-speed inference at 114ms. Extensive experiments demonstrate state-of-the-art performance across standard benchmarks, with pixel error (PE) of 5.43% and corner error (CE) of 4.02% on the LSUN, PE of 7.04% (CE 5.17%) on the Hedau and PE of 4.03% (CE 3.15%) on the Matterport3D-Layout datasets. The framework's combination of geometric awareness and computational efficiency makes it particularly suitable for augmented reality applications and large-scale 3D scene reconstruction tasks.

Existing pre-trained text-to-video diffusion models can generate high-quality videos, but often struggle with misalignment between the generated content and the input text, particularly while composing scenes with multiple objects. To tackle this issue, we propose a straightforward, training-free approach for compositional video generation from text. We introduce Video-ASTAR for test-time aggregation and segregation of attention with a novel centroid loss to enhance alignment, which enables the generation of multiple objects in the scene, modeling the actions and interactions. Additionally, we extend our approach to the Multi-Action video generation setting, where only the specified action should vary across a sequence of prompts. To ensure coherent action transitions, we introduce a novel token-swapping and latent interpolation strategy. Extensive experiments and ablation studies show that our method significantly outperforms baseline methods, generating videos with improved semantic and compositional consistency alongside improved temporal coherence.

Recent interactive image editing methods have made notable progress, yet achieving both precise control and real-time performance remains a challenge. Drag-based methods offer detailed geometric manipulations but suffer from low image fidelity and slow runtime performance, while text-based approaches enhance realism but limit precise and pixel-level control. To overcome these limitations, we introduce Dragonite, an intuitive and efficient framework that seamlessly unifies geometric and semantic manipulation for image editing. Dragonite leverages a Dual Guidance Module that fuses geometric deformation vectors with semantic guidance cues into a joint representation space, ensuring precise manipulation of both content and semantics. By combining a single-step latent optimization mechanism with a enhanced interpolation method, Dragonite achieves efficient interactive image editing while maintaining high precision through integrated geometric and semantic guidance. Extensive evaluations on the DragBench benchmark demonstrate that Dragonite effectively resolves the trade-off between speed and accuracy, enabling real-time, high-fidelity image editing.

Vision Foundation Models, such as \dino[??], have demonstrated remarkable generalization in classification; however, their application to out-of-domain visual regression tasks remains a significant and underexplored challenge. Unlike classification, domain generalization in regression poses distinct challenges: regression produces continuous outputs and is particularly sensitive to high-variance, label-irrelevant factors (e.g., illumination, blur, or contrast). These factors can entangle with task-relevant features and induce spurious correlations. While recent regression methods [??] have shown promise, they often rely on CNN backbones and require the pre-specification of known distractors. This demands significant domain expertise and fails to address spurious correlations that emerge during training. To address these challenges, we propose \proposedapproach, a Latent Distractor Subspace Consistency framework that disentangles intermediate feature representation into task-relevant and latent distractor subspaces, and regularizes the latter under photometric perturbations to suppress spurious correlations while preserving discriminative features during training. Our proposed method, \proposedapproach, is the first to effectively adapt the powerful \dino backbone for domain generalized visual regression. \proposedapproach achieves state-of-the-art results on seven benchmark regression datasets, demonstrating its strong performance in domain generalization for visual regression with percentage improvements of (41.75%, 20.12%, 52.05%, 8.27%, 22.21%, 3.55%) over state-of-the-art \DG regression methods, respectively. Project page is available: \href https://ldsc-iitd.github.io ldsc-iitd.github.io .