WACV 2026 Papers

While representation learning and generative modeling seek to understand visual data, unifying both domains remains unexplored. Recent Unified Self-Supervised Learning (SSL) methods have started to bridge the gap between both paradigms. However, they either extract information from discriminative pretrained models or rely solely on semantic token reconstruction, which requires an external tokenizer during training --- introducing a significant computational overhead. In this work, we introduce Sorcen, a novel Unified SSL framework, incorporating a synergic Contrastive-Reconstruction objective. Our novel Contrastive objective, leverages the generative capabilities of Sorcen and eliminates the need for additional image crops or augmentations during training. Sorcen generates contrastive positive samples, called Echoes, directly in the semantic token space using the reconstruction objective. This on-the-fly Echo generation, enables Sorcen to operate exclusively on precomputed tokens, eliminating the need for an online tokenizer during training. Sorcen significantly reduces the computational overhead by 60.8% compared to token reconstruction SoTA. Extensive experiments on ImageNet-1k demonstrate that Sorcen outperforms the previous Unified SSL SoTA by 0.4%,1.48 FID, 1.76%, and 1.53% on linear probing, unconditional image generation, few-shot learning, and transfer learning, respectively. Additionally, Sorcen establishes as a new single-crop MIM SoTA in linear probing and achieves SoTA performance in unconditional image generation, highlighting significant improvements and breakthroughs in Unified SSL models.

Recent advancements in convolutional neural networks (CNNs) have significantly propelled the field of remote sensing forestry mapping. However, traditional convolution operations exhibit inherent limitations in extracting complex forest features: their fixed receptive fields struggle to accommodate multi-scale forest attributes, and their insufficient focus on background information impairs the overall feature representation. To address these challenges, we propose Similar Convolution (SimConv), which introduces dynamic convolution kernel size selection by modeling feature relationships. SimConv adaptively adjusts the receptive field based on the semantic relevance of input features, enhancing the capture of forestry background information and improving the distinction between target features. Building upon this, we introduce SIMNet, a feature extraction network that integrates SimConv at its core. Experimental results across multiple remote sensing datasets demonstrate that SIMNet outperforms existing methods in terms of feature extraction accuracy. code in https://github.com/WangShiK/SimConv.

Deploying Sentinel-2 satellite-derived bathymetry (SDB)robustly across sites remains challenging. We analyze aSwin-Transformer U-Net (Swin-BathyUNet) to understandhow it infers depth and when its predictions are trustwor-thy. A leave-one-band-out study ranks spectral importanceto the different bands consistent with shallow-water op-tics. We adapt ablation-based CAM to regression (A-CAM-R) and validate faithfulness via a performance-retentiontest: keeping only the top-p% salient pixels while neutral-izing the rest causes large, monotonic RMSE increases,indicating explanations localize causal evidence. Atten-tion ablations show decoder-conditioned cross-attention onskips is the most cost-effective upgrade, improving robust-ness to glint/foam. Cross-region inference (train one site,test another) reveals depth-dependent degradation: MAErises nearly linearly with depth, and bimodal depth dis-tributions exacerbate mid/deep errors. Practical guidancefollows: maintain wide receptive fields, preserve radiomet-ric fidelity in green/blue channels, pre-filter bright high-variance near shore, and pair light target-site fine-tuningwith depth-aware calibration to transfer across regions.

Vision Transformers (ViTs) have achieved state-of-the-art performance in offline video action detection, but their reliance on processing fixed-size clips with full spatio-temporal attention makes them computationally expensive and ill-suited for real-time streaming applications due to massive computational redundancy. This paper introduces a novel framework to adapt these powerful offline models into efficient, online student models through knowledge distillation. We propose two causal, streaming-friendly attention architectures that replace the full self-attention mechanism: (1) a lightweight Temporal Shift Attention that integrates past context with minimal overhead, and (2) a Decomposed Spatial-Temporal Attention that combines intra-frame spatial attention with an LSTM for temporal modeling. Both architectures utilize caching to eliminate redundant operations on a frame-by-frame basis. To maximize knowledge transfer, we introduce an uncertainty-guided distillation process, which formulates the training as a multi-task learning problem. Our resulting models demonstrate significant efficiency gains, achieving up to a 4x improvement in latency and throughput compared to the original offline teacher while ensuring state-of-the-art performance for online methods. Our work provides a practical and effective methodology for deploying high-accuracy transformer models in latency-sensitive, real-world video analysis systems.

Different forms of customized 2D avatars are widely used in gaming applications, virtual communication, education, and content creation. However, existing approaches often fail to capture fine-grained facial expressions and struggle to preserve identity across different expressions. We propose Gen-AFFECT, a novel framework for personalized avatar generation that generates expressive and identity-consistent avatars with a diverse set of facial expressions. Our framework proposes conditioning a multimodal diffusion transformer on an extracted identity-expression representation. This enables identity preservation and representation of a wide range of facial expressions. Gen-AFFECT additionally employs consistent attention at inference for information sharing across the set of generated expressions, enabling the generation process to maintain identity consistency over the array of generated fine-grained expressions. Gen-AFFECT demonstrates superior performance compared to previous state-of-the-art methods on the basis of the accuracy of the generated expressions, the preservation of the identity and the consistency of the target identity across an array of fine-grained facial expressions.

Recent deep learning-based Multi-View Stereo (MVS) approaches, such as MASt3R and VGGT, have shown strong performance in sparse-view 3D reconstruction. However, refining these outputs with 3D Gaussian Splatting (3DGS) remains non-trivial. Excessive positional degrees of freedom (DoFs) in Gaussians often cause instability and geometric artifacts, sometimes distorting geometry to represent texture patterns. To address this issue, we propose GDoFS (Gaussian DoF Separation), a strategy that divides positional DoFs into two categories--image-plane-parallel and ray-aligned--based on their uncertainty. For each category, GDoFS introduces tailored optimization techniques, including bounded offsets for low-uncertainty DoFs and a visibility-guided loss for ray-aligned DoFs. Experiments on standard benchmarks demonstrate that GDoFS effectively mitigates geometric artifacts and produces reconstructions that are both visually coherent and structurally accurate.

We propose a novel generative approach for 3D human pose estimation. 3D human pose estimation poses several key challenges due to the complex geometry of the human body, self-occluding joints, and the requirement for large-scale real-world motion datasets. To address these challenges, we introduce Point2Pose, a framework that effectively models the distribution of human poses conditioned on sequential point cloud and pose history. Specifically, we employ a spatio-temporal point cloud encoder and a pose feature encoder to extract joint-wise features, followed by an attention-based generative regressor. Additionally, we present a large-scale indoor dataset MVPose3D, which contains multiple modalities, including IMU data of non-trivial human motions, dense multi-view point clouds, and RGB images. Experimental results show that the proposed method outperforms the baseline models, demonstrating its superior performance across various datasets.

Unguided sampling in diffusion models is known to generate images of wide variations, albeit with a trade-off in image fidelity. Guided sampling methods, such as classifier guidance (CG) and classifier-free guidance (CFG), focus on sampling in well-learned high-probability regions to generate images of high fidelity, but each has its limitations. CG is computationally expensive due to the costly classifier gradient descent process, while CFG, being gradient-free, is more efficient but compromises class label alignment compared to CG. In this work, we introduce Gradient-Free Classifier Guidance (GFCG), a novel method utilizing a pre-trained classifier solely in inference mode, entirely avoiding gradient calculations. GFCG introduces two innovative mechanisms: (1) adaptive reference class selection, dynamically determining an appropriate undesired reference class; and (2) adaptive guidance scaling, dynamically adjusting the guidance strength based on classifier confidence during sampling. Experiments on both class-conditioned and text-to-image generation demonstrate that the proposed GFCG method consistently improves class prediction accuracy while preserving diversity. We also show that GFCG is complementary to other guided sampling methods. When combined with Autoguidance (ATG), GFCG attains record performance on ImageNet 512x512, with a FD_DINOv2 of 23.39, while maintaining a high Precision of 94.0% compared to ATG's 89.9%.

Reasoning segmentation seeks pixel-accurate masks for targets referenced by complex, often implicit instructions, requiring context-dependent reasoning over the scene. Recent multimodal language models have advanced instruction following segmentation, yet generalization remains limited. The key bottleneck is the high cost of curating diverse, high-quality pixel annotations paired with rich linguistic supervision leading to brittle performance under distribution shift. Therefore, we present CORA, a semi-supervised reasoning segmentation framework that jointly learns from limited labeled data and a large corpus of unlabeled images. CORA introduces three main components: 1) conditional visual instructions that encode spatial and contextual relationships between objects; 2) a noisy pseudo-label filter based on the consistency of Multimodal LLM's outputs across semantically equivalent queries; and 3) a token-level contrastive alignment between labeled and pseudo-labeled samples to enhance feature consistency. These components enable CORA to perform robust reasoning segmentation with minimal supervision, outperforming existing baselines under constrained annotation settings. CORA achieves state-of-the-art results, requiring as few as 100 labeled images on Cityscapes, a benchmark dataset for urban scene understanding, surpassing the baseline by +2.3%. Similarly, CORA improves performance by +2.4% with only 180 labeled images on PanNuke, a histopathology dataset.

We tackle the challenge of robust, in-the-wild imaging using ultra-thin nanophotonic metalens cameras. Meta-lenses, composed of planar arrays of nanoscale scatterers, promise dramatic reductions in size and weight compared to conventional refractive optics. However, severe chromatic aberration, pronounced light scattering, narrow spectral bandwidth, and low light efficiency continue to limit their practical adoption. In this work, we present an end-to-end solution for in-the-wild imaging that pairs a metalens over 12000x thinner than conventional optics with a bespoke multi-image restoration framework optimized for practical metalens cameras. Our method centers on a lightweight convolutional network paired with a memory-efficient burst fusion algorithm that adaptively corrects noise, saturation clipping, and lens-induced distortions across rapid sequences of extremely degraded metalens captures. Extensive experiments on diverse, real-world handheld captures demonstrate that our approach consistently outperforms existing burst-mode and single-image restoration techniques. These results point toward a practical route for deploying metalens-based cameras in everyday imaging applications.

Accurate 6D object pose estimation is vital for robotics, augmented reality, and scene understanding. For seen objects, high accuracy is often attainable via per-object fine-tuning but generalizing to unseen objects remains a challenge. To address this problem, past arts assume access to CAD models at test time and typically follow a multi-stage pipeline to estimate poses: detect and segment the object, propose an initial pose, and then refine it. Under occlusion, however, the early-stage of such pipelines are prone to errors, which can propagate through the sequential processing, and consequently degrade the performance. To remedy this shortcoming, we propose four novel extensions to model-based 6D pose estimation methods: (i) a dynamic non-uniform dense sampling strategy that focuses computation on visible regions, reducing occlusion-induced errors; (ii) a multi-hypothesis inference mechanism that retains several confidence-ranked pose candidates, mitigating brittle single-path failures; (iii) iterative refinement to progressively improve pose accuracy; and (iv) series of occlusion-focused training augmentations that strengthen robustness and generalization. Furthermore, we propose a new weighted by visibility metric for evaluation under occlusion to minimize the bias in the existing protocols. Via extensive empirical evaluations, we show that our proposed approach achieves more than 5% improvement in accuracy on ICBIN and more than 2% on BOP dataset benchmarks, while achieving approximately 3 times faster inference.

Finding information in hour-long videos is a challenging task even for top-performing Vision Language Models (VLMs), as encoding visual content quickly exceeds available context windows. To tackle this challenge, we present FALCONEye, a novel video agent based on a training-free, model-agnostic meta-architecture composed of a VLM and a Large Language Model (LLM). FALCONEye answers open-ended questions using an exploration-based search algorithm guided by calibrated confidence from the VLM's answers. We also introduce FALCON-Bench, a benchmark that extends question answering problem to Video Answer Search--requiring models to return both the answer and its supporting temporal window for open-ended questions in hour-long videos. With just a 7B VLM and a lightweight LLM, FALCONEye outscores all open-source 7B VLMs and comparable agents in FALCONBench. It further demonstrates its generalization capability in MLVU benchmark with shorter videos and different tasks, surpassing GPT-4o on single-detail tasks while slashing inference cost by roughly an order of magnitude.

While the current generation of Vision Language Models (VLMs) has excelled in ideal conditions, their performance drops significantly when exposed to realistic multimodal corruptions, such as blurry images and grammatically incorrect text. Our work addresses this by establishing a novel multimodal corruption and denoising benchmark, VLSRB, with a rich suite of 18 visual and 18 textual corruption functions, to evaluate the system robustness of VLMs. To enhance robustness, we employ: (i) cross-distribution visual denoisers, inspired by the Mixture of Experts (MoE) architecture, and (ii) a prompted zero-shot textual denoiser. Our experiments reveal an overall accuracy gain of up to 5.5%, while revealing the vulnerability of models to specific corruptions and their over-reliance on the textual modality. We envision that the behavioral insights from our benchmark will help in developing robust VLM systems. Our code is available at: https://github.com/farhanishmam/VLMDenoising.

Branches swaying in the breeze, flags rippling in the wind, and boats rocking on the water all show how aerodynamics shape natural motion -- an effect crucial for realism in vision and graphics. In this paper, we present Gaussian Swaying, a surface-based framework for aerodynamic simulation using 3D Gaussians. Unlike mesh-based methods that require costly meshing, or particle-based approaches that rely on discrete positional data, Gaussian Swaying models surfaces continuously with 3D Gaussians, enabling efficient and fine-grained aerodynamic interaction. Our framework unifies simulation and rendering on the same representation: Gaussian patches, which support force computation for dynamics while simultaneously providing normals for lightweight shading. Comprehensive experiments on both synthetic and real-world datasets across multiple metrics demonstrate that Gaussian Swaying achieves state-of-the-art performance and efficiency, offering a scalable approach for realistic aerodynamic scene simulation.

The handling of exclusion in multimodal retrieval remains an underexplored challenge with significant implications for the accuracy and reliability of information retrieval systems. Although existing approaches have advanced multimodal understanding, they typically lack mechanisms to process exclusion explicitly. To address this, we propose a novel model ExclMM that leverages disentangled representations to effectively handle exclusion in multimodal retrieval. Our approach enables precise differentiation between the presence and absence of specific elements in an image, outperforming existing methods. To rigorously evaluate our model, we construct a dataset, ExcluCOCO that pairs exclusion-based queries with ground truth images, sourced from MSCOCO. This dataset serves as a robust benchmark for assessing exclusion comprehension in multimodal contexts. By explicitly incorporating exclusion, our work advances multimodal retrieval by introducing both a model tailored for exclusion-aware retrieval and a benchmark to facilitate future research in this domain.

A conventional camera captures an image of a directly visible scene by measuring the light intensity (and color) arriving at each pixel of its image sensor from a corresponding scene patch. Accordingly, conventional photography treats the measured light as being solely informative of the directly visible scene. Recent research has shown that subtle variations in measured light intensity can carry information about scenes outside the camera's line of sight. A subset of these methods exploits preexisting obstructions---i.e., occluders---which cast barely perceptible yet highly informative soft shadows onto the observed planar surface. Whereas most prior works assume that exploitable occluders are either partly or wholly known or almost planar, a recent work blended a trained diffusion-based sampler to reconstruct the hidden occluding structures in 3D with a transverse 2D radiosity map of all other hidden non-occluding structures. This work proposes a fully trained novel multipath decoding UNet (MDUNet) architecture, in which the multimodal, multipath decoder parallels recent physics-based methods that achieve success by explicitly separating the representations and reconstructions of occluding and non-occluding hidden scene structures. However, by sharing a latent feature representation between the occluding and non-occluding structures, MDUNet couples their reconstruction pathways. Empirical results show that MDUNet improves inference times by over 100x compared to a state-of-the-art diffusion-based method and by 1000x compared to an iterative optimization-based method, while also improving reconstruction quality. In addition, MDUNet is trained solely on simulation data but generalizes to real experimental data, maintaining accuracy and stability even as ambient illumination increases.

Video captioning models have seen notable advancements in recent years, especially with regard to their ability to capture temporal information. While many research efforts have focused on architectural advancements, such as temporal attention mechanisms, there remains a notable gap in understanding how models capture and utilize temporal semantics for effective temporal feature extraction, especially in the context of Advanced Driver Assistance Systems. We propose an automated LiDAR-based captioning procedure that focuses on the temporal dynamics of traffic participants. Our approach uses a rule-based system to extract essential details such as lane position and relative motion from object tracks, followed by a template-based caption generation. Our findings show that training SwinBERT, a video captioning model, using only front camera images and supervised with our template-based captions, specifically designed to encapsulate fine-grained temporal behavior, leads to improved temporal understanding consistently across three datasets. In conclusion, our results clearly demonstrate that integrating LiDAR-based caption supervision significantly enhances temporal understanding, effectively addressing and reducing the inherent visual/static biases prevalent in current state-of-the-art model architectures.

Tables play a key role in conveying structured data across documents. Accurate table detection is crucial for downstream tasks like structure recognition and information extraction. However, current datasets lack diversity in format, language, and layout, limiting real-world generalization. This underscores the need for well-annotated datasets that are multi-lingual, layout-diverse, document-agnostic and format-rich. To address these limitations, we introduce UniTabBank, a large scale, diverse table detection dataset designed to reflect realistic use cases. UniTabBank is characterized by five key attributes: (i) Multi-Lingual --- supporting 28 languages (including Arabic, English, Hindi, etc.); (ii) Multi-Layout --- encompassing both single-column and multi-column documents; (iii) Multi-Type --- covering a wide range of document genres such as annual reports, books, newspapers, and magazines; (iv) Multi-Format --- comprising scanned documents, photographed pages, and PDFs; and finally (v) Scale and Annotation Quality --- consists of 55,443 document page images with 81,179 accurately annotated table instances, offering scale and annotation precision. Additionally, we introduce UniTabDet, a YOLO-based model for table detection, which outperforms state-of-the-arts on eight out of nine table detection benchmarks. Evaluated in a zero-shot setting on four layout analysis benchmarks, UniTabDet also shows strong generalization across diverse documents without additional fine-tuning. The UniTabBank dataset and models are avaialble here.

Multimodal learning has shown promise in medical imaging, combining complementary modalities like images and text. Vision-language models (VLMs) capture rich diagnostic cues but often require large paired datasets and prompt or text-based inference. Their practicality is therefore limited due to annotation cost, privacy, and compute demands. Unpaired external text, like pathology reports, can still provide complementary diagnostic cues if semantically relevant content is retrievable per image. To address this, we introduce CLIP-IT, a novel framework that relies on rich unpaired text reports. Specifically, CLIP-IT uses a CLIP model pre-trained on histology image-text pairs from a separate dataset to retrieve the most relevant unpaired textual report for each image in the downstream unimodal dataset. These reports, sourced from the same disease domain and tissue type, form pseudo-pairs that reflect shared clinical semantics rather than exact alignment. Knowledge from these texts is distilled into the vision model during training, while LoRA-based adaptation mitigates the semantic gap between unaligned modalities. At inference, only the vision model is used, maintaining low overhead while still benefiting from multimodal training without requiring paired data in the downstream dataset. Experiments show that CLIP-IT consistently improves classification accuracy over both unimodal and multimodal CLIP-based baselines in most cases, without requiring paired annotations per dataset or incurring additional inference-time complexity.