WACV 2026 Papers

Sparse-view novel view synthesis is fundamentally ill-posed due to severe geometric ambiguity. Current methods are caught in a trade-off: regressive models are geometrically faithful but incomplete, whereas generative models can complete scenes but often introduce structural inconsistencies. We propose OracleGS, a novel framework that reconciles generative completeness with regressive fidelity for sparse view Gaussian Splatting. Instead of using generative models to patch incomplete reconstructions, our "propose-and-validate" framework first leverages a pre-trained 3D-aware diffusion model to synthesize novel views to propose a complete scene. We then repurpose a multi-view stereo (MVS) model as a 3D-aware oracle to validate the 3D uncertainties of generated views, using its attention maps to reveal regions where the generated views are well-supported by multi-view evidence versus where they fall into regions of high uncertainty due to occlusion, lack of texture, or direct inconsistency. This uncertainty signal directly guides the optimization of a 3D Gaussian Splatting model via an uncertainty-weighted loss. Our approach conditions the powerful generative prior on multi-view geometric evidence, filtering hallucinatory artifacts while preserving plausible completions in under-constrained regions, outperforming state-of-the-art methods on datasets including Mip-NeRF 360 and NeRF Synthetic.

As datasets grow in size, expert-based annotation becomes increasingly impractical, making crowdsourcing a scalable and cost-effective alternative. In crowdsourcing, samples are typically annotated by multiple workers and aggregated via majority voting, a process that overlooks annotator-specific biases and introduces noisy labels that can impair downstream models. Traditional multi-rater methods attempt to model annotator biases (e.g., with transition matrices) but often overfit when faced with many classes or few, noisy annotators. By contrast, Learning with Noisy Labels (LNL) assumes a single noisy label per sample and has demonstrated that robust strategies (e.g., semi-supervised and multi-model learning) usually outperform bias-estimation methods, though these approaches remain underexplored in multi-annotator settings.To bridge this gap, we propose the Reciprocal Teacher-student Learning from Multi-rater Noisy Annotation (RETINA), a framework that integrates LNL techniques into multi-rater learning. RETINA trains annotator-specific models to capture individual labeling patterns and employs a dynamic teacher-student process, where the teacher identifies clean and noisy samples to guide the student. Experiments on synthetic and real-world benchmarks, including our proposed SynMRL benchmark, show that RETINA outperforms existing multi-rater methods, particularly in high-noise, low-annotator, many-class settings.

Synthesizing extrapolated views remains a difficult task, especially in urban driving scenes, where the only reliable sources of data are limited RGB captures and sparse LiDAR points. To address this problem, we present PointmapDiff, a framework for novel view synthesis that utilizes pre-trained 2D diffusion models. Our method leverages point maps (i.e., rasterized 3D scene coordinates) as a conditioning signal, capturing geometric and photometric priors from the reference images to guide the image generation process. With the proposed reference attention layers and ControlNet for point map features, PointmapDiff can generate accurate and consistent results across varying viewpoints while respecting geometric fidelity. Experiments on real-life driving data demonstrate that our method achieves high-quality generation with flexibility over point map conditioning signals (e.g., dense depth map or even sparse LiDAR points) and can be used to distill to 3D representations such as 3D Gaussian Splatting for improving view extrapolation.

We propose to improve the time-sensitive video understanding (TSV) capability of video large language models (Video-LLMs) with grounded objects (GO). We hypothesize that TSV tasks can benefit from GO within frames, which is supported by our preliminary experiments on LITA, a state-of-the-art Video-LLM for reasoning temporal localization. While augmenting prompts with textual descriptions of these object annotations improves the performance of LITA, it also introduces extra token length and susceptibility to the noise in object-level information. To address this, we propose GO-Tokenizer, a lightweight add-on module for Video-LLMs leveraging off-the-shelf object detectors to encode compact object information on the fly. Experimental results demonstrate that pretraining with GO-Tokenizer outperforms the vanilla Video-LLM and its counterpart, utilizing textual descriptions of objects in the prompt. The gain generalizes across different models, datasets, and video understanding tasks, such as reasoning temporal localization and dense captioning.

Robustness audits of deep neural networks (DNN) provide a means to uncover model sensitivities to the challenging real-world imaging conditions that significantly degrade DNN performance in-the-wild. Such conditions are often the result of multiple interacting factors inherent to the environment, sensor, or processing pipeline and may lead to complex image distortions that are not easily categorized. When robustness audits are limited to a set of isolated imaging effects or distortions, the results cannot be (easily) transferred to real-world conditions where image corruptions may be more complex or nuanced. To address this challenge, we present a new alternative robustness auditing method that uses causal inference to measure DNN sensitivities to the factors of the imaging process that cause complex distortions. Our approach uses causal models to explicitly encode assumptions about the domain-relevant factors and their interactions. Then, through extensive experiments on natural and rendered images across multiple vision tasks, we show that our approach reliably estimates causal effects of each factor on DNN performance using only observational domain data. These causal effects directly tie DNN sensitivities to observable properties of the imaging pipeline in the domain of interest towards reducing the risk of unexpected DNN failures when deployed in that domain.

Functional Magnetic Resonance Imaging (fMRI) enables the non-invasive mapping of brain activity during various cognitive processes, and understanding how specific cognitive tasks activate distinct brain regions is crucial for both neuroscience and clinical applications. However, acquiring task-based fMRI (tfMRI) is costly, time-consuming, and often infeasible for individuals unable to perform tasks. Resting-state fMRI (rsfMRI), which is more widely available and does not require task compliance, can instead be leveraged to infer task-related activations. In this work, we propose BrainSparseCNN, a sparse 3D convolutional neural network that exploits the brain's spatial structure to predict tfMRI contrasts from rsfMRI while efficiently processing high-dimensional neuroimaging data. BrainSparseCNN achieves up to 7% higher Pearson correlation than surface-based and dense volumetric competitors, with statistically significant gains (p < 0.01) across all tasks, furthermore improving spatial alignment, subject identification accuracy, and saliency interpretability while maintaining computational efficiency. Overall, BrainSparseCNN provides a more precise, interpretable, and scalable framework for inferring individual functional activation maps from resting-state data, with code publicly available at https://github.com/univanxx/brainsparsecnn.

Dense video captioning aims to interpret and describe all temporally localized events throughout an input video. Recent state-of-the-art methods leverage large language models (LLMs) to provide detailed moment descriptions for video data. However, existing VideoLLMs remain challenging in identifying precise event boundaries in untrimmed videos, causing the generated captions to be not properly grounded. In this paper, we propose TA-Prompting, which enhances VideoLLMs via Temporal Anchors that learn to precisely localize events and prompt the VideoLLMs to perform temporal-aware video event understanding. During inference, in order to properly determine the output caption sequence from an arbitrary number of events presented within a video, we introduce an event coherent sampling strategy to select event captions with sufficient coherence across temporal events and cross-modal similarity with the given video. Through extensive experiments on benchmark datasets, we show that our TA-Prompting is favorable against state-of-the-art VideoLLMs, yielding superior performance on dense video captioning and temporal understanding tasks including moment retrieval and temporalQA.

Few-shot semantic segmentation aims to build robust models that segment unseen objects using only a few labeled examples. Existing FSS approaches, which rely on semantic feature matching, often suffer from Background Bias, Pose-Scale Discrepancy Bias, and the inability to capture fine object details. These limitations hinder their ability to generalize to novel categories, especially in scenarios with high intra-class variability and fine-grained object structures. To overcome these challenges, we propose DOT-Graph, a novel framework that designs CLIP-driven feature Disentanglement and Optimal Transport-based Graph learning for robust few-shot segmentation. We evaluate DOTGraph on PASCAL-5i and COCO-20i, achieving state-of-the-art performance with improvements in various few-shot settings. Our results demonstrate that DOTGraph effectively mitigates background bias, improves feature alignment, and enhances fine-grained segmentation. The code will be released soon.

With the rapid advancement of sophisticated synthetic audio-visual content, e.g., for subtle malicious manipulations, ensuring the perceptual integrity of digital media has become paramount. This work presents a novel approach to temporal localization of deepfakes by leveraging Audio-Visual Speech Representation Reconstruction (AuViRe). Specifically, our approach reconstructs speech representations from one modality (e.g., visual lip movements) based on the other (e.g., audio waveform). This cross-modal reconstruction becomes significantly more challenging, leading to amplified discrepancies, in manipulated regions, thereby providing robust discriminative cues for precise forgery localization. AuViRe outperforms the State-of-the-Art by +8.9 AP@0.95 on LAV-DF, +9.6 AP@0.5 on AV-Deepfake1M, and +5.1 AUC on an in-the-wild experiment. Code will be publicly available upon acceptance.

Assessing student engagement is critical for collaborative learning but remains a challenging task. Existing approaches often rely on controlled laboratory or online settings, which fail to capture the complexity of real-world classrooms. Furthermore, current datasets are scarce and rarely provide both individual- and group-level annotations, limiting the development of robust and generalizable models. To address these gaps, we propose CORE-Net, a multimodal architecture that integrates context modeling to capture group-level dynamics and ordinal supervision to account for the ordinal nature of engagement levels. We also present COLER, a large-scale dataset collected in authentic classroom environments with rich annotations at multiple levels. Experiments demonstrate that CORE- Net achieves 89.63% accuracy and 94.80 QWK, significantly outperforming strong baselines such as BlockGCN and MoViNet. Ablation studies further highlight the critical role of both context modeling and ordinal supervision. Our work establishes a robust and scalable foundation for automated engagement assessment, supporting timely feedback and enhancing the effectiveness of collaborative learning.

In radiation therapy, fiducial markers implanted in a patient's body are tracked using X-ray fluoroscopy to estimate tumor positions.However, flexible markers, such as Gold Anchor^ \textregistered (Naslund Medical AB, Sweden), deform within the body, making conventional template matching challenging.While deep learning offers a promising solution, the extensive collection and annotation of clinical data required for training poses a significant barrier to adoption.To address this, we propose a tracking framework that utilizes a lightweight Siamese CNN trained exclusively on synthetic fluoroscopy images.Our method generates synthetic data simulating diverse marker deformations under low-contrast and high-noise conditions, employs dynamic programming for stable initial detection, and performs real-time tracking with a particle filter.In evaluations using clinical data, our method achieves a tracking accuracy of 0.42 +- 0.12 pixels for prostate cancer cases and 0.97 +- 0.53 pixels for pancreatic cancer cases.This significantly outperforms conventional methods, particularly in challenging low-contrast pancreatic cancer cases.With TensorRT optimization, the framework achieves a processing speed of 3.8 ms/frame.This work presents a practical solution for high-accuracy tracking, reducing data collection costs and facilitating the use of deep learning in clinical applications.

Deep neural networks exploit shortcuts--spurious correlations like laterality markers (spatial) or scanner-specific noise (spectral)--that severely compromise generalization. Many healthcare applications face multiple concurrent shortcuts that are both spatial and spectral, which existing methods struggle to handle. We present SilverLining, an attention-based preprocessing framework that simultaneously identifies and mitigates both spatial and spectral shortcuts without introducing new spurious correlations. Our key insight is that naive removal of shortcut features can itself create new shortcuts, where models learn to exploit the removal patterns as new spurious correlations. We address this through a principled confounder-free correction strategy that maintains consistent preprocessing patterns across all classes in both spatial and frequency domains, preventing new confounders. Extensive experiments demonstrate SilverLining's effectiveness: achieving 0.87 AUC on controlled vision tasks and 0.94 AUC on counter-shortcut medical imaging evaluation where shortcuts are reversed; improving cross-institutional chest X-ray classification from 0.72 to 0.77 AUC; and 0.54 mAP on polyp detection despite natural spurious correlations from surgical overlays. Our data-centric approach provides an effective solution for reducing multiple types of data shortcuts without architectural modifications, creating preprocessed datasets that improve model robustness across both classification and detection tasks. Our codebase is available at https://github.com/theidentity/SilverLining_WACV2026/.

Skill assessment in procedural videos is crucial for the objective evaluation of human performance in settings such as manufacturing and procedural daily tasks. Current research on skill assessment has predominantly focused on sports and lacks large-scale datasets for complex procedural activities. Existing studies typically involve only a limited number of actions, focus on either pairwise assessments (e.g., A is better than B) or on binary labels (e.g., good execution vs needs improvement). In response to these shortcomings, we introduce ProSKILL, the first benchmark dataset for action-level skill assessment in procedural tasks. ProSKILL provides absolute skill assessment annotations, along with pairwise ones. This is enabled by a novel and scalable annotation protocol that allows for the creation of an absolute skill assessment ranking starting from pairwise assessments. This protocol leverages a Swiss Tournament scheme for efficient pairwise comparisons, which are then aggregated into consistent, continuous global scores using an ELO-based rating system. We use our dataset to benchmark the main state-of-the-art skill assessment algorithms, including both ranking-based and pairwise paradigms. The suboptimal results achieved by the current state-of-the-art highlight the challenges and thus the value of ProSKILL in the context of skill assessment for procedural videos.

Conventional point cloud rendering methods are based on the assumption that the complete point cloud is used as input data. However, these methods exhibit limitations in real-world applications, particularly in-stream environments. In real-world scenarios, point cloud acquisition and transmission typically require compression, resulting in degraded point clouds that should be rendered effectively. To address these challenges, we proposed a novel point cloud rendering method through cylindrical projection, optimized for in-stream environments. This approach minimizes distortion by projecting the entire point cloud radially onto a 2D cylindrical plane based on its central axis, enabling grid-based 2D processing. This method provides stable and consistent performance even with sparse or incomplete point clouds. Extensive benchmark experiments have revealed that the proposed framework effectively renders various scenes and objects without requiring additional fine-tuning and demonstrated superior performance and generalization ability in in-stream environments compared with conventional methods.

We address the challenge of generalizing video deblurring models to real-world scenarios, where traditional methods often fail due to a significant domain gap between synthetic and real blur. This work extends the image-to-image translation framework to the more complex domain of video deblurring, introducing a training procedure that effectively bridges this gap. Our method integrates a robust video deblurring backbone with realistic motion priors captured from gimbal-mounted cameras, enabling the model to generalize well across both synthetic and real-world datasets--without requiring paired real-world training data or dataset-specific tuning. Extensive experiments demonstrate that our approach outperforms existing state-of-the-art methods on multiple real-world benchmarks, including datasets never seen during training. Importantly, the generalization capability stems not from the specific architecture, but from the modular training procedure itself, which can be readily applied to other deblurring backbones. This positions our method as a scalable and transferable framework for real-world video deblurring.

Recent advances in generative models, especially diffusion models, have significantly improved image restoration (IR) performance. However, existing problem-agnostic diffusion model-based image restoration (DMIR) methods face challenges in fully leveraging diffusion priors, resulting in suboptimal performance. In this paper, we address the limitations of current problem-agnostic DMIR methods by analyzing their sampling process and providing effective solutions. We introduce EquS, a DMIR method that imposes equivariant information through dual sampling trajectories. To further boost EquS, we propose the Timestep-Aware Schedule (TAS) and introduce EquS^+. TAS prioritizes deterministic steps to enhance certainty and sampling efficiency. Extensive experiments on benchmarks demonstrate that our method is compatible with previous problem-agnostic DMIR methods and significantly boosts their performance without increasing computational costs. Our code is available in https://github.com/FouierL/EquS.

Diffusion models have emerged as the leading approach for style transfer, yet they struggle with photo-realistic transfers, often producing painting-like results or missing detailed stylistic elements. Current methods inadequately address unwanted influence from original content styles and style reference content features. We introduce SCAdapter, a novel technique leveraging CLIP image space to effectively separate and integrate content and style features. Our key innovation systematically extracts pure content from content images and style elements from style references, ensuring authentic transfers. This approach is enhanced through three components: Controllable Style Adaptive Instance Normalization (CSAdaIN) for precise multi-style blending, KVS Injection for targeted style integration, and a style transfer consistency objective maintaining process coherence. Comprehensive experiments demonstrate SCAdapter significantly outperforms state-of-the-art methods in both conventional and diffusion-based baselines. By eliminating DDIM inversion and inference-stage optimization, our method achieves at least 2x faster inference than other diffusion-based approaches, making it both more effective and efficient for practical applications.

Glass is a prevalent material among solid objects in everyday life, yet segmentation methods struggle to distinguish it from opaque materials due to its transparency and reflection. While it is known that human perception relies on boundary and reflective-object features to distinguish glass objects, the existing literature has not yet sufficiently captured both properties when handling transparent objects. Hence, we propose incorporating both of these powerful visual cues via the Boundary Feature Enhancement and Reflection Feature Enhancement modules in a mutually beneficial way. Our proposed framework, TransCues, is a pyramidal transformer encoder-decoder architecture to segment transparent objects. We empirically show that these two modules can be used together effectively, improving overall performance across various benchmark datasets, including glass object semantic segmentation, mirror object semantic segmentation, and generic segmentation datasets. Our method outperforms the state-of-the-art by a large margin, achieving +4.2% mIoU on Trans10K-v2, +5.6% mIoU on MSD, +10.1% mIoU on RGBD-Mirror, +13.1% mIoU on TROSD, and +8.3% mIoU on Stanford2D3D, showing the effectiveness of our method against glass objects.

Alzheimer's disease (AD) is the most common form of dementia and a progressive, irreversible brain disorder that affects millions worldwide. The majority of existing research on AD classification relies on cross-sectional brain magnetic resonance imaging studies, which consider information from a single time point and fail to account for the progressive nature of AD. Longitudinal analysis, however, is crucial for capturing AD evolution and enabling more accurate diagnosis. To address this gap, we propose GeoHSAF, a novel hippocampus-based geometric learning framework for longitudinal AD classification. To overcome the challenge of missing or inconsistent hippocampal shapes across subjects and time points, our framework includes an interpolation module that predicts intermediate shapes, ensuring temporal continuity. We evaluate the effectiveness of GeoHSAF on three public longitudinal AD datasets: ADNI, OASIS, and AIBL, and benchmark its performance against existing approaches. GeoHSAF achieves new state-of-the-art results on binary classification tasks (AD vs. NC), while also demonstrating strong performance on more challenging triple-class classification tasks (AD vs. NC vs. MCI). Our work is fully reproducible, and all code is available at: https://github.com/ayodejimb/GeoHSAF