WACV 2026 Papers

The prevalence of user-generated content (UGC) on platforms such as YouTube and TikTok has rendered no-reference (NR) perceptual video quality assessment (VQA) vital for optimizing video delivery. Nonetheless, the characteristics of non-professional acquisition and the subsequent transcoding of UGC video on sharing platforms present significant challenges for NR-VQA. Although NR-VQA models attempt to infer mean opinion scores (MOS), their modeling of subjective scores for compressed content remains limited due to the absence of fine-grained perceptual annotations of artifact types. To address these challenges, we propose CAMP-VQA, a novel NR-VQA framework that exploits the semantic understanding capabilities of large vision-language models. Our approach introduces a quality-aware prompting mechanism that integrates video metadata (e.g., resolution, frame rate, bitrate) with key fragments extracted from inter-frame variations to guide the BLIP-2 pretraining approach in generating fine-grained quality captions. A unified architecture has been designed to model perceptual quality across three dimensions: semantic alignment, temporal characteristics, and spatial characteristics. These multimodal features are extracted and fused, then regressed to video quality scores. Extensive experiments on a wide variety of UGC datasets demonstrate that our model consistently outperforms existing NR-VQA methods, achieving improved accuracy without the need for costly manual fine-grained annotations. Our method achieves the best performance in terms of average rank and linear correlation (SRCC: 0.928, PLCC: 0.938) compared to state-of-the-art methods. The source code and trained models, along with a user-friendly demo, are available at: https://github.com/xinyiW915/CAMP-VQA.

We introduce SSMRadNet, the first multi-scale State Space Model (SSM) based detector for Frequency Modulated Continuous Wave (FMCW) radar that sequentially processes raw ADC samples through two SSMs. One SSM learns a chirp-wise feature by sequentially processing samples from all receiver channels within one chirp, and a second SSM learns a representation of a frame by sequentially processing chirp-wise features. The latent representations of a radar frame are decoded to perform segmentation and detection tasks. Comprehensive evaluations on the RADIal dataset show SSMRadNet has 10-33x fewer parameters and 60-88x less computation (GFLOPs) while being 3.7x faster than state-of-the-art transformer and convolution-based radar detectors at competitive performance for segmentation tasks.

Video diffusion models (VDMs) have demonstrated remarkable capabilities in text-to-video (T2V) generation. Despite their success, VDMs still suffer from degraded image quality and flickering artifacts. To address these issues, some approaches have introduced preference learning to exploit human feedback to enhance the video generation. However, these methods primarily adopt the routine in the image domain without an in-depth investigation into video-specific preference optimization. In this paper, we reexamine the design of the video preference learning from two key aspects: feedback source and feedback tuning methodology, and present OnlineVPO, a more efficient preference learning framework tailored specifically for VDMs. On the feedback source, we found that the image-level reward model commonly used in existing methods fails to provide a human-aligned video preference signal due to the modality gap. In contrast, video quality assessment (VQA) models show superior alignment with human perception of video quality. Building on this insight, we propose leveraging VQA models as a proxy of humans to provide more modality-aligned feedback for VDMs. Regarding the preference tuning methodology, we introduce an online DPO algorithm tailored for VDMs. It not only enjoys the benefits of superior scalability in optimizing videos with higher resolution and longer duration compared with the existing method, but also mitigates the insufficient optimization issue caused by off-policy learning via online preference generation and curriculum preference update designs. Extensive experiments on the open-source video-diffusion model demonstrate OnlineVPO as a simple yet effective and, more importantly, scalable preference learning algorithm for video diffusion models.

Implicit neural representation (INR) has emerged as a new data representation for compressing videos and now shows on par performance with the conventional codecs. The next quest in the field is to make INR scalable for its practical use. Existing works realize this by utilizing small INR models to scale for long and high-resolution video, which achieves better encoding and decoding speeds. However, they fail to fully exploit the temporal nature of video data when encoding it into multiple separate INRs across time, which leads to suboptimal compression efficiency. In this work, we propose NerVast, a new encoding scheme for video INR, that improves compression efficiency while still enjoying the low computation and transfer costs of small INR models. When a video is represented in separate INR segments, NerVast effectively reduces the total volume required for representation by sharing the parameters between models during encoding. Without expensive training, NerVast selects the most efficient parameters to share. Then it jointly trains both shared and non-shared parameters in a way that minimizes the quality drop imposed by sharing. While maintaining real-time decoding speed (> 30 fps), NerVast provides better compression (39.9% reduction in parameters on average) compared to the compute-efficient INR models. In other words, NerVast is better in encoding quality (1.57 dB higher in PSNR) with the same bitrate.

Advancements in 3D curve skeletonization are accelerating progress across a wide range of applications. However, developing robust skeletonization algorithms that capture intricate object details remains challenging. Skeletonization via Local Separators (LS) offers an efficient graph-based approach but suffers from representation inaccuracies due to its discrete nature. To address this, we introduce CSCD, a novel framework for Curve Skeletonization in the Continuous Domain, generalizing LS to manifolds. Specifically, we present two realizations: CSCD-M for meshes and CSCD-PC for point clouds. CSCD-M leverages the intrinsic triangulation of a mesh for resilience to noise and improved topological preservation, while CSCD-PC employs tufted Laplacians for enhanced robustness. To our knowledge, CSCD-M is the first intrinsic method for curve skeletonization. Our results show CSCD-M matches LS performance across diverse meshes and outperforms LS (TOG'21) on benchmarks like Thingi10k dataset. CSCD-PC qualitatively outperforms CoverageAxis++ (Eurographics'24) and EPCS (CAG'23). Finally, we demonstrate the efficacy of CSCD in a few downstream tasks: object classification, shape segmentation, and identifying handles, tunnels, and constrictions in objects. Website: https://cscd-skel.pages.dev

With the increasing demand for histopathological specimen examination and diagnostic reporting, Multiple Instance Learning (MIL) has received heightened research focus as a viable solution for AI-centric diagnostic aid. Recently, to improve its performance and make it work more like a pathologist, several MIL approaches based on the use of multiple-resolution images have been proposed, delivering often higher performance than those that use single-resolution images. Despite impressive recent developments of multiple-resolution MIL, previous approaches only focus on improving performance, thereby lacking research on well-calibrated MIL that clinical experts can rely on for trustworthy diagnostic results. In this study, we propose Uncertainty-Focused Calibrated MIL (UFC-MIL), which more closely mimics the pathologists' examination behaviors while providing calibrated diagnostic predictions, using multiple images with different resolutions. UFC-MIL includes a novel patch-wise loss that learns the latent patterns of instances and expresses their uncertainty for classification. Also, the attention-based architecture with a neighbor patch aggregation module collects features for the classifier. In addition, aggregated predictions are calibrated through patch-level uncertainty without requiring multiple iterative inferences, which is a key practical advantage. Against challenging public datasets, UFC-MIL shows superior performance in model calibration while achieving classification accuracy comparable to that of state-of-the-art methods.

Recent models like OpenAI's O1 and DeepSeek's R1, which utilize test-time scaling techniques, have demonstrated remarkable improvements in reasoning capabilities. We anticipate that in the near future, multimodal models will also experience significant breakthroughs in multimodal reasoning. This will require some highly challenging and specialized evaluations.As one of the most crucial real-world applications of multimodal models, visual agents require complex and comprehensive capabilities such as spatial planning and vision-in-the-chain type reasoning. These capabilities are currently lacking in existing multimodal benchmarks. In this paper, we introduce **MageBench**, a **M**ultimodal reasoning benchmark built upon light-weight **AGE**nt environments that pose significant reasoning challenges and hold substantial practical value. The results show that only a few product-level models are better than random acting, and all of them are far inferior to human level. We analyze and summarize their errors and capability gaps in visual planning.Furthermore, we found that rule-based RL can significantly boost visual reasoning capabilities. This highlights that our benchmark could serve as a valuable testing ground for the emerging field of agentic RL research.

Accurate state representation is critical for effective procedure planning from visual inputs. Existing methods typically rely on sentence-level natural language descriptions to represent states. However, such representations are often ambiguous and fail to capture fine-grained object interactions, leading to errors in complex scenarios. To overcome these limitations, we propose a Hybrid State Representation (HSR), inspired by human-like reasoning, that models procedural states with both structural precision and contextual clarity. HSR integrates two complementary modalities: (1) Semantic State Graphs (SSGs), which explicitly encode objects, attributes, and their relations, and (2) contextual Question-Answer (QA) pairs, which act as semantic probes to disambiguate critical state transitions. We further design a heterogeneous encoder to fuse these components and introduce a visual-state alignment objective to ground the hybrid representation in the visual context. Extensive experiments on the CrossTask, COIN, and NIV benchmarks demonstrate that our method establishes a new state-of-the-art, achieving significant gains on the strict Success Rate (SR) metric. Ablation studies confirm that both the structural (SSG) and contextual (QA) components of HSR are essential for the observed performance gains.

The increasing need for robust digital signature verification systems has amplified interest in realistic online signature generation to counter digital forgeries. In this work, we propose a novel Decision Transformer based framework that learns Reinforcement Learning output to generate diverse online signatures. Departing from traditional RL approaches that rely on policy gradients or value function estimation, we formulate signature generation as a sequence-modeling problem. Our framework addresses varied free-form signature styles, demonstrating adaptability across linguistic and stylistic variations. Initially, an RL model generates signature trajectories, which are then fed to Decision Transformer, employing an autoregressive sequence modeling approach. To further personalize the generated signatures, we introduce a Q-learning-based module that produces user-specific variations while mitigating noise. By operating in an offline reinforcement learning setting, the proposed method reduces the dependency on extensive online interactions, improving scalability. Experimental results on a publicly available online signature dataset in multiple linguistic script styles show that our approach significantly outperforms traditional generative methods in terms of realism, variability, and mimicry accuracy. These results highlight the potential of Decision Transformers for structured sequence generation tasks beyond their conventional domains.

Synthesizing high-fidelity frozen 3D scenes from monocular Mannequin-Challenge (MC) videos is a unique problem distinct from standard dynamic scene reconstruction. Instead of focusing on modeling motion, our goal is to create a frozen scene while strategically preserving subtle dynamics to enable user-controlled instant selection. To achieve this, we introduce a novel application of dynamic Gaussian splatting: the scene is modeled dynamically, which retains nearby temporal variation, and a static scene is rendered by fixing the model's time parameter. However, under this usage, monocular capture with sparse temporal supervision introduces artifacts like ghosting and blur for Gaussians that become unobserved or occluded at weakly supervised timestamps. We propose Splannequin, an architecture-agnostic regularization that detects two states of Gaussian primitives, hidden and defective, and applies temporal anchoring. Under predominantly forward camera motion, hidden states are anchored to their well-observed past states, while defective states are anchored to future states with stronger supervision. Our method integrates into existing dynamic Gaussian pipelines via simple loss terms, requires no architectural changes, and adds zero inference overhead. This results in markedly improved visual quality, enabling high-fidelity, user-selectable frozen-time renderings, validated by a 96% user preference.

Future trajectories of neighboring traffic agents have a significant influence on the path planning and decision-making of autonomous vehicles. While trajectory forecasting is a well-studied field, research mainly focuses on snapshot-based prediction, where each scenario is treated independently of its global temporal context. However, real-world autonomous driving systems need to operate in a continuous setting, requiring real-time processing of data streams with low latency and consistent predictions over successive timesteps. We leverage this continuous setting to propose a lightweight yet highly accurate streaming-based trajectory forecasting approach. We integrate valuable information from previous predictions with a novel endpoint-aware modeling scheme. Our temporal context propagation uses the trajectory endpoints of the previous forecasts as anchors to extract targeted scenario context encodings. Our approach efficiently guides its scene encoder to extract highly relevant context information without needing refinement iterations or segment-wise decoding. Our experiments highlight that our approach effectively relays information across consecutive timesteps. Unlike methods using multi-stage refinement processing, our approach significantly reduces inference latency, making it well-suited for real-world deployment.We achieve state-of-the-art streaming trajectory prediction results on the Argoverse 2 multi-agent and single-agent benchmarks, while requiring substantially fewer resources.

Generating SVGs from text prompts is a challenging vision task, requiring diverse yet realistic depictions of the seen as well as unseen entities. However, existing research has been mostly limited to generating single-object rather than comprehensive scenes comprising multiple elements. In response, CraftSVG, introduces an end-to-end framework for creating SVGs depicting entire scenes from textual descriptions. Utilizing a pre-trained LLM for layout generation from text prompts via iterative in-context learning, CraftSVG introduces a technique for producing masked latent in specified bounding boxes for accurate object placement. It introduces a fusion mechanism for integrating attention maps and employs a diffusion U-Net for coherent composition, speeding up the stroke initialization. Recognizing the importance of abstract SVGs in communication, we incorporated an MLP-based mechanism to simplify the resulting SVGs, with alignment and perceptual loss with differential rendering and opacity modulation to maximize the similarity. CraftSVG outperforms previous methods in abstraction, recognizability, and detail, as demonstrated by its performance metrics: CLIP-T: 0.5013, Cosine Similarity: 0.7091, and Aesthetic: 7.0779, among others.

Diffusion models demonstrate remarkable capabilities in capturing complex data distributions and have achieved compelling results in many generative tasks. While they have recently been extended to dense prediction tasks such as depth estimation and surface normal prediction, their full potential in this area remains underexplored. As target signal maps and input images are pixel-wise aligned, the conventional noise-to-data generation paradigm is inefficient, and input images can serve as a more informative prior compared to pure noise. Diffusion bridge models, which support data-to-data generation between two general data distributions, offer a promising alternative, but they typically fail to exploit the rich visual priors embedded in large pretrained foundation models. To address these limitations, we integrate diffusion bridge formulation with structured visual priors and introduce DPBridge, the first latent diffusion bridge framework for dense prediction tasks. To resolve the incompatibility between diffusion bridge models and pretrained diffusion backbones, we propose (1) a tractable reverse transition kernel for the diffusion bridge process, enabling maximum likelihood training scheme; (2) finetuning strategies including distribution-aligned normalization and image consistency loss. Experiments across extensive benchmarks validate that our method consistently achieves superior performance, demonstrating its effectiveness and generalization capability under different scenarios.

Generating expressive, fine-grained human motion from text remains a formidable challenge, particularly when aiming for high fidelity without incurring excessive computational cost. Existing methods often rely on complex, multi-stage pipelines with slow inference and large memory footprints, hindering real-time deployment. To address these limitations, we introduce MoSCo, a simple autoregressive text-to-motion framework that discretizes motion into part-level token sequences and models temporal dynamics via Delta-based training strategy --i.e., predicting the motion difference from the previous time step--before fusing these tokens with textual embeddings through our Part-Aware Coordinator(PAO) and generating the full sequence with a single, lightweight transformer decoder. MoSCo sets a new milestone in text-to-motion inference speed--achieving an AITS of just 0.002s(vs 0.03s), over an order of magnitude faster than all prior methods--while maintaining a compact model footprint and delivering highly realistic motions (FID 0.085), making real-time, high-quality generation practical.

Imperfect reconstructions arising from occlusions, shadows, reflections, and other factors during 3D scanning often result in incomplete sections of the scanned object, with missing parts scattered randomly across its surface. We introduce an uncertainty-aware signed distance field (SDF) latent transformer that leverages uncertainty to identify and reconstruct missing parts based on the global shape of the incomplete scanned object and the immediate neighborhood of the affected regions. To our knowledge, we are the first to utilize uncertainties for SDF shape completion in the latent space. Our model has been trained on the entire Objaverse 1.0 dataset and demonstrates that our uncertainty-aware SDF completion method significantly outperforms previous works both numerically and visually. Code will be published at github.com/cgtuebingen/ua3dscancomp.

Unsupervised anomaly detection in medical images is an important yet challenging task due to the diversity of possible anomalies and the practical impossibility of collecting comprehensively annotated datasets. In this paper, we propose a modernized autoencoder-based framework, the Q-Former Autoencoder, that leverages state-of-the-art pretrained vision foundation models for medical anomaly detection. Instead of training encoders from scratch, we directly utilize frozen foundation models as feature extractors, enabling rich, multi-stage, high-level representations without domain-specific fine-tuning. We introduce the Q-Former architecture as the bottleneck, which enables us to control the length of the reconstruction sequence, while efficiently aggregating multi-scale features. Additionally, we incorporate a perceptual loss computed using features from a the trained Masked Autoencoder, guiding the reconstruction towards semantically meaningful structures. Our framework is evaluated on four diverse medical anomaly detection benchmarks - BraTS2021, RESC, RSNA and LiverCT, achieving state-of-the-art results. Our results highlight the potential of foundation model encoders, pretrained on natural images, to generalize effectively to medical image analysis tasks without further fine-tuning. Code and models will be released upon acceptance.

Implicit neural representation (INR) methods for video compression have recently achieved visual quality and compression ratios that are competitive with traditional pipelines. However, due to the need for per-sample network training, the encoding speeds of these methods are too slow for practical adoption. We develop a library to allow us to disentangle and review the components of methods from the NeRV family, reframing their performance in terms of not only size-quality trade-offs, but also impacts on training time. We uncover principles for effective video INR design and propose a state-of-the-art configuration of these components, Rabbit NeRV (RNeRV). When all methods are given equal training time (equivalent to 300 NeRV epochs) for 7 different UVG videos at 1080p, RNeRV achieves +1.27% PSNR on average compared to the best-performing alternative for each video in our NeRV library. We then tackle the encoding speed issue head-on by investigating the viability of hyper-networks, which predict INR weights from video inputs, to disentangle training from encoding to allow for real-time encoding. We propose masking the weights of the predicted INR during training to allow for variable, higher quality compression, resulting in 1.7% improvements to both PSNR and MS-SSIM at 0.037 bpp on the UCF-101 dataset, and we increase hyper-network parameters by 0.4% for 2.5%/2.7% improvements to PSNR/MS-SSIM with equal bpp and similar speeds.

Detecting small, recreational vessels in coastal environments remains a persistent challenge due to complex backgrounds, dynamic lighting conditions, and the scarcity of annotated data for non-commercial maritime traffic. Despite their socio-economic significance, recreational boats are underrepresented in existing datasets and are poorly detected by standard object detectors, particularly in open-vocabulary scenarios. To address this gap, we present RampWatch, an in-the-wild dataset curated from surveillance footage at multiple boat ramps. RampWatch provides instance annotations across seven categories of recreational vessels, captured under diverse weather, lighting, and occlusion conditions. To benchmark detection in this domain, we introduce YOLO-TG, a novel detection framework that augments YOLOv11 with a text encoder for open-vocabulary recognition and a self-attention module for enhanced spatial reasoning. YOLO-TG adopts a dual-stream design: visual features are extracted via a hierarchical YOLO backbone, while semantic embeddings from natural language prompts are encoded by a frozen language encoder. These modalities are fused via lightweight cross-modal attention, enabling text-guided detection without retraining. YOLO-TG achieves a +12% relative improvement in mAP@50-95 over strong YOLOv11 baselines on RampWatch, and demonstrates robust cross-domain generalization, with gains of +22% on the Singapore Maritime Dataset and +4.3% on the Split Port Ship Classification Dataset. These results highlight the effectiveness of cross-modal grounding and domain-specific datasets for advancing open-world maritime surveillance.

When multiple videos of a scene are taken from differing viewpoints without precise synchronization, it can be difficult to temporally align them after the fact. Often the metadata or audio needed to do so is missing or inaccurate. But human motion in such videos can provide a strong signal for identifying matching time points across videos, through analysis of pose and movement. In this work, we leverage view-invariant human pose features to synchronize videos. Unlike previous human pose-based alignment techniques, our method can align videos containing multiple people without performing tracking or re-identification across views. We achieve this by aggregating pose information from multiple people into a single frame descriptor. This also enables fast \mathcal O (n\log n) search for the optimal alignment. This simple but effective strategy leads to major and consistent improvements over existing human-based and visual feature temporal alignment techniques.