Theo Gevers

Qi Bi, Shaodi You, Theo Gevers

Domain-generalized urban-scene semantic segmentation (USSS) aims to learn generalized semantic predictions across diverse urban-scene styles. Unlike domain gap challenges, USSS is unique in that the semantic categories are often similar in different urban scenes, while the styles can vary significantly due to changes in urban landscapes, weather conditions, lighting, and other factors. Existing approaches typically rely on convolutional neural networks (CNNs) to learn the content of urban scenes. In this paper, we propose a Content-enhanced Mask TransFormer (CMFormer) for domain-generalized USSS. The main idea is to enhance the focus of the fundamental component, the mask attention mechanism, in Transformer segmentation models on content information. To achieve this, we introduce a novel content-enhanced mask attention mechanism. It learns mask queries from both the image feature and its down-sampled counterpart, as lower-resolution image features usually contain more robust content information and are less sensitive to style variations. These features are fused into a Transformer decoder and integrated into a multi-resolution content-enhanced mask attention learning scheme. Extensive experiments conducted on various domain-generalized urban-scene segmentation datasets demonstrate that the proposed CMFormer significantly outperforms existing CNN-based methods for domain-generalized semantic segmentation, achieving improvements of up to 14.00\% in terms of mIoU (mean intersection over union). The source code is publicly available at \url{https://github.com/BiQiWHU/CMFormer}.

Partha Das, Sezer Karaoglu, Arjan Gijsenij et al.

Intrinsic image decomposition (IID) is an under-constrained problem. Therefore, traditional approaches use hand crafted priors to constrain the problem. However, these constraints are limited when coping with complex scenes. Deep learning-based approaches learn these constraints implicitly through the data, but they often suffer from dataset biases (due to not being able to include all possible imaging conditions). In this paper, a combination of the two is proposed. Component specific priors like semantics and invariant features are exploited to obtain semantically and physically plausible reflectance transitions. These transitions are used to steer a progressive CNN with implicit homogeneity constraints to decompose reflectance and shading maps. An ablation study is conducted showing that the use of the proposed priors and progressive CNN increase the IID performance. State of the art performance on both our proposed dataset and the standard real-world IIW dataset shows the effectiveness of the proposed method. Code is made available at https://github.com/Morpheus3000/SIGNet

Qi Ma, Yue Li, Bin Ren et al.

3D Gaussian Splatting (3DGS) has become the de facto method of 3D representation in many vision tasks. This calls for the 3D understanding directly in this representation space. To facilitate the research in this direction, we first build ShapeSplat, a large-scale dataset of 3DGS using the commonly used ShapeNet, ModelNet and Objaverse datasets. Our dataset ShapeSplat consists of 206K objects spanning over 87 unique categories, whose labels are in accordance with the respective datasets. The creation of this dataset utilized the compute equivalent of 3.8 GPU years on a TITAN XP GPU. We utilize our dataset for unsupervised pretraining and supervised finetuning for classification and segmentation tasks. To this end, we introduce Gaussian-MAE, which highlights the unique benefits of representation learning from Gaussian parameters. Through exhaustive experiments, we provide several valuable insights. In particular, we show that (1) the distribution of the optimized GS centroids significantly differs from the uniformly sampled point cloud (used for initialization) counterpart; (2) this change in distribution results in degradation in classification but improvement in segmentation tasks when using only the centroids; (3) to leverage additional Gaussian parameters, we propose Gaussian feature grouping in a normalized feature space, along with splats pooling layer, offering a tailored solution to effectively group and embed similar Gaussians, which leads to notable improvement in finetuning tasks.

Dimitrije Antić, Georgios Paschalidis, Shashank Tripathi et al.

Recovering 3D object pose and shape from a single image is a challenging and ill-posed problem. This is due to strong (self-)occlusions, depth ambiguities, the vast intra- and inter-class shape variance, and the lack of 3D ground truth for natural images. Existing deep-network methods are trained on synthetic datasets to predict 3D shapes, so they often struggle generalizing to real-world images. Moreover, they lack an explicit feedback loop for refining noisy estimates, and primarily focus on geometry without directly considering pixel alignment. To tackle these limitations, we develop a novel render-and-compare optimization framework, called SDFit. This has three key innovations: First, it uses a learned category-specific and morphable signed-distance-function (mSDF) model, and fits this to an image by iteratively refining both 3D pose and shape. The mSDF robustifies inference by constraining the search on the manifold of valid shapes, while allowing for arbitrary shape topologies. Second, SDFit retrieves an initial 3D shape that likely matches the image, by exploiting foundational models for efficient look-up into 3D shape databases. Third, SDFit initializes pose by establishing rich 2D-3D correspondences between the image and the mSDF through foundational features. We evaluate SDFit on three image datasets, i.e., Pix3D, Pascal3D+, and COMIC. SDFit performs on par with SotA feed-forward networks for unoccluded images and common poses, but is uniquely robust to occlusions and uncommon poses. Moreover, it requires no retraining for unseen images. Thus, SDFit contributes new insights for generalizing in the wild. Code is available at https://anticdimi.github.io/sdfit.

Rick Groenendijk, Leo Dorst, Theo Gevers

Pooling is essentially an operation from the field of Mathematical Morphology, with max pooling as a limited special case. The more general setting of MorphPooling greatly extends the tool set for building neural networks. In addition to pooling operations, encoder-decoder networks used for pixel-level predictions also require unpooling. It is common to combine unpooling with convolution or deconvolution for up-sampling. However, using its morphological properties, unpooling can be generalised and improved. Extensive experimentation on two tasks and three large-scale datasets shows that morphological pooling and unpooling lead to improved predictive performance at much reduced parameter counts.

Başak Melis Öcal, Maxim Tatarchenko, Sezer Karaoglu et al.

Designing high-quality indoor 3D scenes is important in many practical applications, such as room planning or game development. Conventionally, this has been a time-consuming process which requires both artistic skill and familiarity with professional software, making it hardly accessible for layman users. However, recent advances in generative AI have established solid foundation for democratizing 3D design. In this paper, we propose a pioneering approach for text-based 3D room design. Given a prompt in natural language describing the object placement in the room, our method produces a high-quality 3D scene corresponding to it. With an additional text prompt the users can change the appearance of the entire scene or of individual objects in it. Built using in-context learning, CAD model retrieval and 3D-Gaussian-Splatting-based stylization, our turnkey pipeline produces state-of-the-art 3D scenes, while being easy to use even for novices. Our project page is available at https://sceneteller.github.io/.

Xiaoyan Xing, Konrad Groh, Sezer Karaoglu et al.

The purpose of intrinsic decomposition is to separate an image into its albedo (reflective properties) and shading components (illumination properties). This is challenging because it's an ill-posed problem. Conventional approaches primarily concentrate on 2D imagery and fail to fully exploit the capabilities of 3D data representation. 3D point clouds offer a more comprehensive format for representing scenes, as they combine geometric and color information effectively. To this end, in this paper, we introduce Point Intrinsic Net (PoInt-Net), which leverages 3D point cloud data to concurrently estimate albedo and shading maps. The merits of PoInt-Net include the following aspects. First, the model is efficient, achieving consistent performance across point clouds of any size with training only required on small-scale point clouds. Second, it exhibits remarkable robustness; even when trained exclusively on datasets comprising individual objects, PoInt-Net demonstrates strong generalization to unseen objects and scenes. Third, it delivers superior accuracy over conventional 2D approaches, demonstrating enhanced performance across various metrics on different datasets. (Code Released)

Xiaoyan Xing, Vincent Tao Hu, Jan Hendrik Metzen et al.

This paper introduces a novel approach to illumination manipulation in diffusion models, addressing the gap in conditional image generation with a focus on lighting conditions. We conceptualize the diffusion model as a black-box image render and strategically decompose its energy function in alignment with the image formation model. Our method effectively separates and controls illumination-related properties during the generative process. It generates images with realistic illumination effects, including cast shadow, soft shadow, and inter-reflections. Remarkably, it achieves this without the necessity for learning intrinsic decomposition, finding directions in latent space, or undergoing additional training with new datasets.

Ruihong Yin, Sezer Karaoglu, Theo Gevers

In addition to color and textural information, geometry provides important cues for 3D scene reconstruction. However, current reconstruction methods only include geometry at the feature level thus not fully exploiting the geometric information. In contrast, this paper proposes a novel geometry integration mechanism for 3D scene reconstruction. Our approach incorporates 3D geometry at three levels, i.e. feature learning, feature fusion, and network supervision. First, geometry-guided feature learning encodes geometric priors to contain view-dependent information. Second, a geometry-guided adaptive feature fusion is introduced which utilizes the geometric priors as a guidance to adaptively generate weights for multiple views. Third, at the supervision level, taking the consistency between 2D and 3D normals into account, a consistent 3D normal loss is designed to add local constraints. Large-scale experiments are conducted on the ScanNet dataset, showing that volumetric methods with our geometry integration mechanism outperform state-of-the-art methods quantitatively as well as qualitatively. Volumetric methods with ours also show good generalization on the 7-Scenes and TUM RGB-D datasets.

Yue Li, Qi Ma, Runyi Yang et al.

While 3DGS has emerged as a high-fidelity scene representation, encoding rich, general-purpose features directly from its primitives remains under-explored. We address this gap by introducing Chorus, a multi-teacher pretraining framework that learns a holistic feed-forward 3D Gaussian Splatting (3DGS) scene encoder by distilling complementary signals from 2D foundation models. Chorus employs a shared 3D encoder and teacher-specific projectors to learn from language-aligned, generalist, and object-aware teachers, encouraging a shared embedding space that captures signals from high-level semantics to fine-grained structure. We evaluate Chorus on a wide range of tasks: open-vocabulary semantic and instance segmentation, linear and decoder probing, as well as data-efficient supervision. Besides 3DGS, we also test Chorus on several benchmarks that only support point clouds by pretraining a variant using only Gaussians' centers, colors, estimated normals as inputs. Interestingly, this encoder shows strong transfer and outperforms the point clouds baseline while using 39.9 times fewer training scenes. Finally, we propose a render-and-distill adaptation that facilitates out-of-domain finetuning. Our code and model will be released upon publication.

Ruihong Yin, Yunlu Chen, Sezer Karaoglu et al.

Existing methods in neural scene reconstruction utilize the Signed Distance Function (SDF) to model the density function. However, in indoor scenes, the density computed from the SDF for a sampled point may not consistently reflect its real importance in volume rendering, often due to the influence of neighboring objects. To tackle this issue, our work proposes a novel approach for indoor scene reconstruction, which instead parameterizes the density function with the Signed Ray Distance Function (SRDF). Firstly, the SRDF is predicted by the network and transformed to a ray-conditioned density function for volume rendering. We argue that the ray-specific SRDF only considers the surface along the camera ray, from which the derived density function is more consistent to the real occupancy than that from the SDF. Secondly, although SRDF and SDF represent different aspects of scene geometries, their values should share the same sign indicating the underlying spatial occupancy. Therefore, this work introduces a SRDF-SDF consistency loss to constrain the signs of the SRDF and SDF outputs. Thirdly, this work proposes a self-supervised visibility task, introducing the physical visibility geometry to the reconstruction task. The visibility task combines prior from predicted SRDF and SDF as pseudo labels, and contributes to generating more accurate 3D geometry. Our method implemented with different representations has been validated on indoor datasets, achieving improved performance in both reconstruction and view synthesis.

Osman Ülger, Yu Wang, Ysbrand Galama et al.

Humans have a remarkable ability to perceive and reason about the world around them by understanding the relationships between objects. In this paper, we investigate the effectiveness of using such relationships for object detection and instance segmentation. To this end, we propose a Relational Prior-based Feature Enhancement Model (RP-FEM), a graph transformer that enhances object proposal features using relational priors. The proposed architecture operates on top of scene graphs obtained from initial proposals and aims to concurrently learn relational context modeling for object detection and instance segmentation. Experimental evaluations on COCO show that the utilization of scene graphs, augmented with relational priors, offer benefits for object detection and instance segmentation. RP-FEM demonstrates its capacity to suppress improbable class predictions within the image while also preventing the model from generating duplicate predictions, leading to improvements over the baseline model on which it is built.

Başak Melis Öcal, Maxim Tatarchenko, Sezer Karaoglu et al.

Point cloud completion aims to recover the complete 3D shape of an object from partial observations. While approaches relying on synthetic shape priors achieved promising results in this domain, their applicability and generalizability to real-world data are still limited. To tackle this problem, we propose a self-supervised framework, namely RealDiff, that formulates point cloud completion as a conditional generation problem directly on real-world measurements. To better deal with noisy observations without resorting to training on synthetic data, we leverage additional geometric cues. Specifically, RealDiff simulates a diffusion process at the missing object parts while conditioning the generation on the partial input to address the multimodal nature of the task. We further regularize the training by matching object silhouettes and depth maps, predicted by our method, with the externally estimated ones. Experimental results show that our method consistently outperforms state-of-the-art methods in real-world point cloud completion.

Qi Zhang, Denis Rozumny, Francesco Girlanda et al.

We propose Unblur-SLAM, a novel RGB SLAM pipeline for sharp 3D reconstruction from blurred image inputs. In contrast to previous work, our approach is able to handle different types of blur and demonstrates state-of-the-art performance in the presence of both motion blur and defocus blur. Moreover, we adjust the computation effort with the amount of blur in the input image. As a first stage, our method uses a feed-forward image deblurring model for which we propose a suitable training scheme that can improve both tracking and mapping modules. Frames that are successfully deblurred by the feed-forward network obtain refined poses and depth through local-global multi-view optimization and loop closure. Frames that fail the first stage deblurring are directly modeled through the global 3DGS representation and an additional blur network to model multiple blurred sub-frames and simulate the blur formation process in 3D space, thereby learning sharp details and refined sub-frame poses. Experiments on several real-world datasets demonstrate consistent improvements in both pose estimation and sharp reconstruction results of geometry and texture.

Wenhan Chen, Sezer Karaoglu, Theo Gevers

Recent advances in diffusion-based generation techniques enable AI models to produce highly realistic videos, heightening the need for reliable detection mechanisms. However, existing detection methods provide only limited exploration of the 3D geometric patterns present in generated videos. In this paper, we use vanishing points as an explicit representation of 3D geometry patterns, revealing fundamental discrepancies in geometric consistency between real and AI-generated videos. We introduce Grab-3D, a geometry-aware transformer framework for detecting AI-generated videos based on 3D geometric temporal consistency. To enable reliable evaluation, we construct an AI-generated video dataset of static scenes, allowing stable 3D geometric feature extraction. We propose a geometry-aware transformer equipped with geometric positional encoding, temporal-geometric attention, and an EMA-based geometric classifier head to explicitly inject 3D geometric awareness into temporal modeling. Experiments demonstrate that Grab-3D significantly outperforms state-of-the-art detectors, achieving robust cross-domain generalization to unseen generators.

Anil S. Baslamisli, Theo Gevers

Intrinsic image decomposition aims to factorize an image into albedo (reflectance) and shading (illumination) sub-components. Being ill-posed and under-constrained, it is a very challenging computer vision problem. There are infinite pairs of reflectance and shading images that can reconstruct the same input. To address the problem, Intrinsic Images in the Wild provides an optimization framework based on a dense conditional random field (CRF) formulation that considers long-range material relations. We improve upon their model by introducing illumination invariant image descriptors: color ratios. The color ratios and the reflectance intrinsic are both invariant to illumination and thus are highly correlated. Through detailed experiments, we provide ways to inject the color ratios into the dense CRF optimization. Our approach is physics-based, learning-free and leads to more accurate and robust reflectance decompositions.

Melis Ocal, Xiaoyan Xing, Yue Li et al.

3D stylization is central to game development, virtual reality, and digital arts, where the demand for diverse assets calls for scalable methods that support fast, high-fidelity manipulation. Existing text-to-3D stylization methods typically distill from 2D image editors, requiring time-intensive per-asset optimization and exhibiting multi-view inconsistency due to the limitations of current text-to-image models, which makes them impractical for large-scale production. In this paper, we introduce GaussianBlender, a pioneering feed-forward framework for text-driven 3D stylization that performs edits instantly at inference. Our method learns structured, disentangled latent spaces with controlled information sharing for geometry and appearance from spatially-grouped 3D Gaussians. A latent diffusion model then applies text-conditioned edits on these learned representations. Comprehensive evaluations show that GaussianBlender not only delivers instant, high-fidelity, geometry-preserving, multi-view consistent stylization, but also surpasses methods that require per-instance test-time optimization - unlocking practical, democratized 3D stylization at scale.

Vladimir Yugay, Thies Kersten, Luca Carlone et al.

Mapping systems with novel view synthesis (NVS) capabilities, most notably 3D Gaussian Splatting (3DGS), are widely used in computer vision, as well as in various applications, including augmented reality, robotics, and autonomous driving. However, many current approaches are limited to static scenes. While recent works have begun addressing short-term dynamics (motion within the camera's view), long-term dynamics (the scene evolving through changes out of view) remain less explored. To overcome this limitation, we introduce a dynamic scene adaptation mechanism to continuously update 3DGS to reflect the latest changes. Since maintaining consistency remains challenging due to stale observations disrupting the reconstruction process, we further propose a novel keyframe management mechanism that discards outdated observations while preserving as much information as possible. We thoroughly evaluate Gaussian Mapping for Evolving Scenes (GaME) on both synthetic and real-world datasets, achieving a 29.7% improvement in PSNR and a 3 times improvement in L1 depth error over the most competitive baseline.

Weijie Wei, Fatemeh Karimi Nejadasl, Theo Gevers et al.

The scarcity of annotated data in LiDAR point cloud understanding hinders effective representation learning. Consequently, scholars have been actively investigating efficacious self-supervised pre-training paradigms. Nevertheless, temporal information, which is inherent in the LiDAR point cloud sequence, is consistently disregarded. To better utilize this property, we propose an effective pre-training strategy, namely Temporal Masked Auto-Encoders (T-MAE), which takes as input temporally adjacent frames and learns temporal dependency. A SiamWCA backbone, containing a Siamese encoder and a windowed cross-attention (WCA) module, is established for the two-frame input. Considering that the movement of an ego-vehicle alters the view of the same instance, temporal modeling also serves as a robust and natural data augmentation, enhancing the comprehension of target objects. SiamWCA is a powerful architecture but heavily relies on annotated data. Our T-MAE pre-training strategy alleviates its demand for annotated data. Comprehensive experiments demonstrate that T-MAE achieves the best performance on both Waymo and ONCE datasets among competitive self-supervised approaches. Codes will be released at https://github.com/codename1995/T-MAE

Weijie Wei, Martin R. Oswald, Fatemeh Karimi Nejadasl et al.

In this paper, we focus on semantic segmentation method for point clouds of urban scenes. Our fundamental concept revolves around the collaborative utilization of diverse scene representations to benefit from different context information and network architectures. To this end, the proposed network architecture, called APNet, is split into two branches: a point cloud branch and an aerial image branch which input is generated from a point cloud. To leverage the different properties of each branch, we employ a geometry-aware fusion module that is learned to combine the results of each branch. Additional separate losses for each branch avoid that one branch dominates the results, ensure the best performance for each branch individually and explicitly define the input domain of the fusion network assuring it only performs data fusion. Our experiments demonstrate that the fusion output consistently outperforms the individual network branches and that APNet achieves state-of-the-art performance of 65.2 mIoU on the SensatUrban dataset. Upon acceptance, the source code will be made accessible.

Mengjiao Ma, Qi Ma, Yue Li et al.

3D Gaussian Splatting (3DGS) serves as a highly performant and efficient encoding of scene geometry, appearance, and semantics. Moreover, grounding language in 3D scenes has proven to be an effective strategy for 3D scene understanding. Current Language Gaussian Splatting line of work fall into three main groups: (i) per-scene optimization-based, (ii) per-scene optimization-free, and (iii) generalizable approach. However, most of them are evaluated only on rendered 2D views of a handful of scenes and viewpoints close to the training views, limiting ability and insight into holistic 3D understanding. To address this gap, we propose the first large-scale benchmark that systematically assesses these three groups of methods directly in 3D space, evaluating on 1060 scenes across three indoor datasets and one outdoor dataset. Benchmark results demonstrate a clear advantage of the generalizable paradigm, particularly in relaxing the scene-specific limitation, enabling fast feed-forward inference on novel scenes, and achieving superior segmentation performance. We further introduce GaussianWorld-49K a carefully curated 3DGS dataset comprising around 49K diverse indoor and outdoor scenes obtained from multiple sources, with which we demonstrate the generalizable approach could harness strong data priors. Our codes, benchmark, and datasets will be made public to accelerate research in generalizable 3DGS scene understanding.

Rick Groenendijk, Leo Dorst, Theo Gevers

Signals from different modalities each have their own combination algebra which affects their sampling processing. RGB is mostly linear; depth is a geometric signal following the operations of mathematical morphology. If a network obtaining RGB-D input has both kinds of operators available in its layers, it should be able to give effective output with fewer parameters. In this paper, morphological elements in conjunction with more familiar linear modules are used to construct a mixed linear-morphological network called HaarNet. This is the first large-scale linear-morphological hybrid, evaluated on a set of sizeable real-world datasets. In the network, morphological Haar sampling is applied to both feature channels in several layers, which splits extreme values and high-frequency information such that both can be processed to improve both modalities. Moreover, morphologically parameterised ReLU is used, and morphologically-sound up-sampling is applied to obtain a full-resolution output. Experiments show that HaarNet is competitive with a state-of-the-art CNN, implying that morphological networks are a promising research direction for geometry-based learning tasks.

Vladimir Yugay, Yue Li, Theo Gevers et al.

We present a dense simultaneous localization and mapping (SLAM) method that uses 3D Gaussians as a scene representation. Our approach enables interactive-time reconstruction and photo-realistic rendering from real-world single-camera RGBD videos. To this end, we propose a novel effective strategy for seeding new Gaussians for newly explored areas and their effective online optimization that is independent of the scene size and thus scalable to larger scenes. This is achieved by organizing the scene into sub-maps which are independently optimized and do not need to be kept in memory. We further accomplish frame-to-model camera tracking by minimizing photometric and geometric losses between the input and rendered frames. The Gaussian representation allows for high-quality photo-realistic real-time rendering of real-world scenes. Evaluation on synthetic and real-world datasets demonstrates competitive or superior performance in mapping, tracking, and rendering compared to existing neural dense SLAM methods.

Weijie Wei, Osman Ülger, Fatemeh Karimi Nejadasl et al.

Open-Vocabulary Segmentation (OVS) methods offer promising capabilities in detecting unseen object categories, but the category must be known and needs to be provided by a human, either via a text prompt or pre-labeled datasets, thus limiting their scalability. We propose 3D-AVS, a method for Auto-Vocabulary Segmentation of 3D point clouds for which the vocabulary is unknown and auto-generated for each input at runtime, thus eliminating the human in the loop and typically providing a substantially larger vocabulary for richer annotations. 3D-AVS first recognizes semantic entities from image or point cloud data and then segments all points with the automatically generated vocabulary. Our method incorporates both image-based and point-based recognition, enhancing robustness under challenging lighting conditions where geometric information from LiDAR is especially valuable. Our point-based recognition features a Sparse Masked Attention Pooling (SMAP) module to enrich the diversity of recognized objects. To address the challenges of evaluating unknown vocabularies and avoid annotation biases from label synonyms, hierarchies, or semantic overlaps, we introduce the annotation-free Text-Point Semantic Similarity (TPSS) metric for assessing generated vocabulary quality. Our evaluations on nuScenes and ScanNet200 demonstrate 3D-AVS's ability to generate semantic classes with accurate point-wise segmentations. Codes will be released at https://github.com/ozzyou/3D-AVS

Hoàng-Ân Lê, Tushar Nimbhorkar, Thomas Mensink et al.

There hardly exists any large-scale datasets with dense optical flow of non-rigid motion from real-world imagery as of today. The reason lies mainly in the required setup to derive ground truth optical flows: a series of images with known camera poses along its trajectory, and an accurate 3D model from a textured scene. Human annotation is not only too tedious for large databases, it can simply hardly contribute to accurate optical flow. To circumvent the need for manual annotation, we propose a framework to automatically generate optical flow from real-world videos. The method extracts and matches objects from video frames to compute initial constraints, and applies a deformation over the objects of interest to obtain dense optical flow fields. We propose several ways to augment the optical flow variations. Extensive experimental results show that training on our automatically generated optical flow outperforms methods that are trained on rigid synthetic data using FlowNet-S, LiteFlowNet, PWC-Net, and RAFT. Datasets and implementation of our optical flow generation framework are released at https://github.com/lhoangan/arap_flow

Yue Li, Qi Ma, Runyi Yang et al.

Recognizing arbitrary or previously unseen categories is essential for comprehensive real-world 3D scene understanding. Currently, all existing methods rely on 2D or textual modalities during training or together at inference. This highlights the clear absence of a model capable of processing 3D data alone for learning semantics end-to-end, along with the necessary data to train such a model. Meanwhile, 3D Gaussian Splatting (3DGS) has emerged as the de facto standard for 3D scene representation across various vision tasks. However, effectively integrating semantic reasoning into 3DGS in a generalizable manner remains an open challenge. To address these limitations, we introduce SceneSplat, to our knowledge the first large-scale 3D indoor scene understanding approach that operates natively on 3DGS. Furthermore, we propose a self-supervised learning scheme that unlocks rich 3D feature learning from unlabeled scenes. To power the proposed methods, we introduce SceneSplat-7K, the first large-scale 3DGS dataset for indoor scenes, comprising 7916 scenes derived from seven established datasets, such as ScanNet and Matterport3D. Generating SceneSplat-7K required computational resources equivalent to 150 GPU days on an L4 GPU, enabling standardized benchmarking for 3DGS-based reasoning for indoor scenes. Our exhaustive experiments on SceneSplat-7K demonstrate the significant benefit of the proposed method over the established baselines.

Vladimir Yugay, Theo Gevers, Martin R. Oswald

Simultaneous localization and mapping (SLAM) systems with novel view synthesis capabilities are widely used in computer vision, with applications in augmented reality, robotics, and autonomous driving. However, existing approaches are limited to single-agent operation. Recent work has addressed this problem using a distributed neural scene representation. Unfortunately, existing methods are slow, cannot accurately render real-world data, are restricted to two agents, and have limited tracking accuracy. In contrast, we propose a rigidly deformable 3D Gaussian-based scene representation that dramatically speeds up the system. However, improving tracking accuracy and reconstructing a globally consistent map from multiple agents remains challenging due to trajectory drift and discrepancies across agents' observations. Therefore, we propose new tracking and map-merging mechanisms and integrate loop closure in the Gaussian-based SLAM pipeline. We evaluate MAGiC-SLAM on synthetic and real-world datasets and find it more accurate and faster than the state of the art.

Xiaoyan Xing, Konrad Groh, Sezer Karaoglu et al.

We introduce LumiNet, a novel architecture that leverages generative models and latent intrinsic representations for effective lighting transfer. Given a source image and a target lighting image, LumiNet synthesizes a relit version of the source scene that captures the target's lighting. Our approach makes two key contributions: a data curation strategy from the StyleGAN-based relighting model for our training, and a modified diffusion-based ControlNet that processes both latent intrinsic properties from the source image and latent extrinsic properties from the target image. We further improve lighting transfer through a learned adaptor (MLP) that injects the target's latent extrinsic properties via cross-attention and fine-tuning. Unlike traditional ControlNet, which generates images with conditional maps from a single scene, LumiNet processes latent representations from two different images - preserving geometry and albedo from the source while transferring lighting characteristics from the target. Experiments demonstrate that our method successfully transfers complex lighting phenomena including specular highlights and indirect illumination across scenes with varying spatial layouts and materials, outperforming existing approaches on challenging indoor scenes using only images as input.

Ruihong Yin, Vladimir Yugay, Yue Li et al.

The field of novel view synthesis from images has seen rapid advancements with the introduction of Neural Radiance Fields (NeRF) and more recently with 3D Gaussian Splatting. Gaussian Splatting became widely adopted due to its efficiency and ability to render novel views accurately. While Gaussian Splatting performs well when a sufficient amount of training images are available, its unstructured explicit representation tends to overfit in scenarios with sparse input images, resulting in poor rendering performance. To address this, we present a 3D Gaussian-based novel view synthesis method using sparse input images that can accurately render the scene from the viewpoints not covered by the training images. We propose a multi-stage training scheme with matching-based consistency constraints imposed on the novel views without relying on pre-trained depth estimation or diffusion models. This is achieved by using the matches of the available training images to supervise the generation of the novel views sampled between the training frames with color, geometry, and semantic losses. In addition, we introduce a locality preserving regularization for 3D Gaussians which removes rendering artifacts by preserving the local color structure of the scene. Evaluation on synthetic and real-world datasets demonstrates competitive or superior performance of our method in few-shot novel view synthesis compared to existing state-of-the-art methods.

Qi Bi, Shaodi You, Theo Gevers

Images from outdoor scenes may be taken under various weather conditions. It is well studied that weather impacts the performance of computer vision algorithms and needs to be handled properly. However, existing algorithms model weather condition as a discrete status and estimate it using multi-label classification. The fact is that, physically, specifically in meteorology, weather are modeled as a continuous and transitional status. Instead of directly implementing hard classification as existing multi-weather classification methods do, we consider the physical formulation of multi-weather conditions and model the impact of physical-related parameter on learning from the image appearance. In this paper, we start with solid revisit of the physics definition of weather and how it can be described as a continuous machine learning and computer vision task. Namely, we propose to model the weather uncertainty, where the level of probability and co-existence of multiple weather conditions are both considered. A Gaussian mixture model is used to encapsulate the weather uncertainty and a uncertainty-aware multi-weather learning scheme is proposed based on prior-posterior learning. A novel multi-weather co-presence estimation transformer (MeFormer) is proposed. In addition, a new multi-weather co-presence estimation (MePe) dataset, along with 14 fine-grained weather categories and 16,078 samples, is proposed to benchmark both conventional multi-label weather classification task and multi-weather co-presence estimation task. Large scale experiments show that the proposed method achieves state-of-the-art performance and substantial generalization capabilities on both the conventional multi-label weather classification task and the proposed multi-weather co-presence estimation task. Besides, modeling weather uncertainty also benefits adverse-weather semantic segmentation.

Xiaoyan Xing, Xiao Zhang, Sezer Karaoglu et al.

Image-to-image relighting requires representations that disentangle scene properties from illumination. Recent methods rely on latent intrinsic representations but remain under-constrained and often fail on challenging materials such as metal and glass. A natural hypothesis is that stronger pretrained visual priors should resolve these failures. We find the opposite: features from top-performing semantic encoders often degrade relighting quality, revealing a fundamental trade-off between semantic abstraction and photometric fidelity. We study this trade-off and introduce Augmented Latent Intrinsics (ALI), which balances semantic context and dense photometric structure by fusing features from a pixel-aligned visual encoder into a latent-intrinsic framework, together with a self-supervised refinement strategy to mitigate the scarcity of paired real-world data. Trained only on unlabeled real-world image pairs and paired with a dense, pixel-aligned visual prior, ALI achieves strong improvements in relighting, with the largest gains on complex, specular materials. Project page: https:\\augmented-latent-intrinsics.github.io

Ruihong Yin, Xuepeng Shi, Oleksandr Bailo et al.

Recent perception-generalist approaches based on language models have achieved state-of-the-art results across diverse tasks, including 3D scene layout estimation, via unified architecture and interface. However, these approaches rely on autoregressive next-token prediction, which is inherently slow. In this work, we introduce Fast SceneScript, a novel structured language model for accurate and efficient 3D scene layout estimation. Our method employs multi-token prediction (MTP) to reduce the number of autoregressive iterations and significantly accelerate inference. While MTP improves speed, unreliable token predictions can significantly reduce accuracy. To filter out unreliable tokens, we adapt self-speculative decoding (SSD) for structured language models and introduce confidence-guided decoding (CGD) with an improved scoring mechanism for token reliability. Furthermore, we design a parameter-efficient mechanism that reduces the parameter overhead of MTP. Extensive experiments on the ASE and Structured3D benchmarks demonstrate that Fast SceneScript can generate up to 9 tokens per decoder inference step without compromising accuracy, while adding only $\sim7.5\%$ additional parameters.

Yanni Ma, Hao Liu, Yulan Guo et al.

3D scene graph prediction aims to abstract complex 3D environments into structured graphs consisting of objects and their pairwise relationships. Existing approaches typically adopt object-centric graph neural networks, where relation edge features are iteratively updated by aggregating messages from connected object nodes. However, this design inherently restricts relation representations to pairwise object context, making it difficult to capture high-order relational dependencies that are essential for accurate relation prediction. To address this limitation, we propose a Link-guided Edge-centric relational reasoning framework with Object-aware fusion, namely LEO, which enables progressive reasoning from relation-level context to object-level understanding. Specifically, LEO first predicts potential links between object pairs to suppress irrelevant edges, and then transforms the original scene graph into a line graph where each relation is treated as a node. A line graph neural network is applied to perform edge-centric relational reasoning to capture inter-relation context. The enriched relation features are subsequently integrated into the original object-centric graph to enhance object-level reasoning and improve relation prediction. Our framework is model-agnostic and can be integrated with any existing object-centric method. Experiments on the 3DSSG dataset with two competitive baselines show consistent improvements, highlighting the effectiveness of our edge-to-object reasoning paradigm.

Vlardimir Yugay, Duy-Kien Nguyen, Theo Gevers et al.

Despite the rapid development of large 3D models, classical optimization-based approaches dominate the field of visual odometry (VO). Thus, current approaches to VO heavily rely on camera parameters and many handcrafted components, most of which involve complex bundle adjustment and feature-matching processes. Although disregarded in the literature, we find it problematic in terms of both (1) speed, that performs bundle adjustment requires a significant amount of time, and (2) scalability, as hand-crafted components struggle to learn from large-scale training data. In this work, we introduce a simple yet efficient architecture, Visual Odometry Transformer (VoT), that formulates monocular visual odometry as a direct relative pose regression problem. Our approach streamlines the monocular visual odometry pipeline in an end-to-end manner, effectively eliminating the need for handcrafted components such as bundle adjustment, feature matching, or camera calibration. We show that VoT is up to 4 times faster than traditional approaches, yet with competitive or better performance. Compared to recent 3D foundation models, VoT runs 10 times faster with strong scaling behavior in terms of both model sizes and training data. Moreover, VoT generalizes well in both low-data regimes and previously unseen scenarios, reducing the gap between optimization-based and end-to-end approaches.

Partha Das, Sezer Karaoglu, Theo Gevers

Intrinsic image decomposition is the process of recovering the image formation components (reflectance and shading) from an image. Previous methods employ either explicit priors to constrain the problem or implicit constraints as formulated by their losses (deep learning). These methods can be negatively influenced by strong illumination conditions causing shading-reflectance leakages. Therefore, in this paper, an end-to-end edge-driven hybrid CNN approach is proposed for intrinsic image decomposition. Edges correspond to illumination invariant gradients. To handle hard negative illumination transitions, a hierarchical approach is taken including global and local refinement layers. We make use of attention layers to further strengthen the learning process. An extensive ablation study and large scale experiments are conducted showing that it is beneficial for edge-driven hybrid IID networks to make use of illumination invariant descriptors and that separating global and local cues helps in improving the performance of the network. Finally, it is shown that the proposed method obtains state of the art performance and is able to generalise well to real world images. The project page with pretrained models, finetuned models and network code can be found at https://ivi.fnwi.uva.nl/cv/pienet/.

Partha Das, Yang Liu, Sezer Karaoglu et al.

In this paper, the aim is multi-illumination color constancy. However, most of the existing color constancy methods are designed for single light sources. Furthermore, datasets for learning multiple illumination color constancy are largely missing. We propose a seed (physics driven) based multi-illumination color constancy method. GANs are exploited to model the illumination estimation problem as an image-to-image domain translation problem. Additionally, a novel multi-illumination data augmentation method is proposed. Experiments on single and multi-illumination datasets show that our methods outperform sota methods.

Hoang-An Le, Thomas Mensink, Partha Das et al.

Multimodal large-scale datasets for outdoor scenes are mostly designed for urban driving problems. The scenes are highly structured and semantically different from scenarios seen in nature-centered scenes such as gardens or parks. To promote machine learning methods for nature-oriented applications, such as agriculture and gardening, we propose the multimodal synthetic dataset for Enclosed garDEN scenes (EDEN). The dataset features more than 300K images captured from more than 100 garden models. Each image is annotated with various low/high-level vision modalities, including semantic segmentation, depth, surface normals, intrinsic colors, and optical flow. Experimental results on the state-of-the-art methods for semantic segmentation and monocular depth prediction, two important tasks in computer vision, show positive impact of pre-training deep networks on our dataset for unstructured natural scenes. The dataset and related materials will be available at https://lhoangan.github.io/eden.

Ipek Ganiyusufoglu, L. Minh Ngô, Nedko Savov et al.

For deepfake detection, video-level detectors have not been explored as extensively as image-level detectors, which do not exploit temporal data. In this paper, we empirically show that existing approaches on image and sequence classifiers generalize poorly to new manipulation techniques. To this end, we propose spatio-temporal features, modeled by 3D CNNs, to extend the generalization capabilities to detect new sorts of deepfake videos. We show that spatial features learn distinct deepfake-method-specific attributes, while spatio-temporal features capture shared attributes between deepfake methods. We provide an in-depth analysis of how the sequential and spatio-temporal video encoders are utilizing temporal information using DFDC dataset arXiv:2006.07397. Thus, we unravel that our approach captures local spatio-temporal relations and inconsistencies in the deepfake videos while existing sequence encoders are indifferent to it. Through large scale experiments conducted on the FaceForensics++ arXiv:1901.08971 and Deeper Forensics arXiv:2001.03024 datasets, we show that our approach outperforms existing methods in terms of generalization capabilities.

Hoang-An Le, Thomas Mensink, Partha Das et al.

In this paper the argument is made that for true novel view synthesis of objects, where the object can be synthesized from any viewpoint, an explicit 3D shape representation isdesired. Our method estimates point clouds to capture the geometry of the object, which can be freely rotated into the desired view and then projected into a new image. This image, however, is sparse by nature and hence this coarse view is used as the input of an image completion network to obtain the dense target view. The point cloud is obtained using the predicted pixel-wise depth map, estimated from a single RGB input image,combined with the camera intrinsics. By using forward warping and backward warpingbetween the input view and the target view, the network can be trained end-to-end without supervision on depth. The benefit of using point clouds as an explicit 3D shape for novel view synthesis is experimentally validated on the 3D ShapeNet benchmark. Source code and data will be available at https://lhoangan.github.io/pc4novis/.

Rick Groenendijk, Sezer Karaoglu, Theo Gevers et al.

Many interesting tasks in machine learning and computer vision are learned by optimising an objective function defined as a weighted linear combination of multiple losses. The final performance is sensitive to choosing the correct (relative) weights for these losses. Finding a good set of weights is often done by adopting them into the set of hyper-parameters, which are set using an extensive grid search. This is computationally expensive. In this paper, we propose a weighting scheme based on the coefficient of variations and set the weights based on properties observed while training the model. The proposed method incorporates a measure of uncertainty to balance the losses, and as a result the loss weights evolve during training without requiring another (learning based) optimisation. In contrast to many loss weighting methods in literature, we focus on single-task multi-loss problems, such as monocular depth estimation and semantic segmentation, and show that multi-task approaches for loss weighting do not work on those single-tasks. The validity of the approach is shown empirically for depth estimation and semantic segmentation on multiple datasets.

Anil S. Baslamisli, Yang Liu, Sezer Karaoglu et al.

We investigate the use of photometric invariance and deep learning to compute intrinsic images (albedo and shading). We propose albedo and shading gradient descriptors which are derived from physics-based models. Using the descriptors, albedo transitions are masked out and an initial sparse shading map is calculated directly from the corresponding RGB image gradients in a learning-free unsupervised manner. Then, an optimization method is proposed to reconstruct the full dense shading map. Finally, we integrate the generated shading map into a novel deep learning framework to refine it and also to predict corresponding albedo image to achieve intrinsic image decomposition. By doing so, we are the first to directly address the texture and intensity ambiguity problems of the shading estimations. Large scale experiments show that our approach steered by physics-based invariant descriptors achieve superior results on MIT Intrinsics, NIR-RGB Intrinsics, Multi-Illuminant Intrinsic Images, Spectral Intrinsic Images, As Realistic As Possible, and competitive results on Intrinsic Images in the Wild datasets while achieving state-of-the-art shading estimations.

Wei Wang, Shaodi You, Sezer Karaoglu et al.

Kinship verification is a well-explored task: identifying whether or not two persons are kin. In contrast, kinship identification has been largely ignored so far. Kinship identification aims to further identify the particular type of kinship. An extension to kinship verification run short to properly obtain identification, because existing verification networks are individually trained on specific kinships and do not consider the context between different kinship types. Also, existing kinship verification datasets have biased positive-negative distributions which are different than real-world distributions. To this end, we propose a novel kinship identification approach based on joint training of kinship verification ensembles and classification modules. We propose to rebalance the training dataset to become more realistic. Large scale experiments demonstrate the appealing performance on kinship identification. The experiments further show significant performance improvement of kinship verification when trained on the same dataset with more realistic distributions.

Anil S. Baslamisli, Partha Das, Hoang-An Le et al.

In general, intrinsic image decomposition algorithms interpret shading as one unified component including all photometric effects. As shading transitions are generally smoother than reflectance (albedo) changes, these methods may fail in distinguishing strong photometric effects from reflectance variations. Therefore, in this paper, we propose to decompose the shading component into direct (illumination) and indirect shading (ambient light and shadows) subcomponents. The aim is to distinguish strong photometric effects from reflectance variations. An end-to-end deep convolutional neural network (ShadingNet) is proposed that operates in a fine-to-coarse manner with a specialized fusion and refinement unit exploiting the fine-grained shading model. It is designed to learn specific reflectance cues separated from specific photometric effects to analyze the disentanglement capability. A large-scale dataset of scene-level synthetic images of outdoor natural environments is provided with fine-grained intrinsic image ground-truths. Large scale experiments show that our approach using fine-grained shading decompositions outperforms state-of-the-art algorithms utilizing unified shading on NED, MPI Sintel, GTA V, IIW, MIT Intrinsic Images, 3DRMS and SRD datasets.

Rick Groenendijk, Sezer Karaoglu, Theo Gevers et al.

In this paper we address the benefit of adding adversarial training to the task of monocular depth estimation. A model can be trained in a self-supervised setting on stereo pairs of images, where depth (disparities) are an intermediate result in a right-to-left image reconstruction pipeline. For the quality of the image reconstruction and disparity prediction, a combination of different losses is used, including L1 image reconstruction losses and left-right disparity smoothness. These are local pixel-wise losses, while depth prediction requires global consistency. Therefore, we extend the self-supervised network to become a Generative Adversarial Network (GAN), by including a discriminator which should tell apart reconstructed (fake) images from real images. We evaluate Vanilla GANs, LSGANs and Wasserstein GANs in combination with different pixel-wise reconstruction losses. Based on extensive experimental evaluation, we conclude that adversarial training is beneficial if and only if the reconstruction loss is not too constrained. Even though adversarial training seems promising because it promotes global consistency, non-adversarial training outperforms (or is on par with) any method trained with a GAN when a constrained reconstruction loss is used in combination with batch normalisation. Based on the insights of our experimental evaluation we obtain state-of-the art monocular depth estimation results by using batch normalisation and different output scales.

Minh Ngô, Burak Mandira, Selim Fırat Yılmaz et al.

Lies and deception are common phenomena in society, both in our private and professional lives. However, humans are notoriously bad at accurate deception detection. Based on the literature, human accuracy of distinguishing between lies and truthful statements is 54% on average, in other words it is slightly better than a random guess. While people do not much care about this issue, in high-stakes situations such as interrogations for series crimes and for evaluating the testimonies in court cases, accurate deception detection methods are highly desirable. To achieve a reliable, covert, and non-invasive deception detection, we propose a novel method that jointly extracts reliable low- and high-level facial features namely, 3D facial geometry, skin reflectance, expression, head pose, and scene illumination in a video sequence. Then these features are modeled using a Recurrent Neural Network to learn temporal characteristics of deceptive and honest behavior. We evaluate the proposed method on the Real-Life Trial (RLT) dataset that contains high-stake deceptive and honest videos recorded in courtrooms. Our results show that the proposed method (with an accuracy of 72.8%) improves the state of the art as well as outperforming the use of manually coded facial attributes 67.6%) in deception detection.

Jian Han, Sezer Karaoglu, Hoang-An Le et al.

In this paper, we provide a synthetic data generator methodology with fully controlled, multifaceted variations based on a new 3D face dataset (3DU-Face). We customized synthetic datasets to address specific types of variations (scale, pose, occlusion, blur, etc.), and systematically investigate the influence of different variations on face detection performances. We examine whether and how these factors contribute to better face detection performances. We validate our synthetic data augmentation for different face detectors (Faster RCNN, SSH and HR) on various face datasets (MAFA, UFDD and Wider Face).

Partha Das, Anil S. Baslamisli, Yang Liu et al.

In this paper, we formulate the color constancy task as an image-to-image translation problem using GANs. By conducting a large set of experiments on different datasets, an experimental survey is provided on the use of different types of GANs to solve for color constancy i.e. CC-GANs (Color Constancy GANs). Based on the experimental review, recommendations are given for the design of CC-GAN architectures based on different criteria, circumstances and datasets.

Wei Zeng, Sezer Karaoglu, Theo Gevers

In this paper, we propose a pipeline to generate 3D point cloud of an object from a single-view RGB image. Most previous work predict the 3D point coordinates from single RGB images directly. We decompose this problem into depth estimation from single images and point cloud completion from partial point clouds. Our method sequentially predicts the depth maps from images and then infers the complete 3D object point clouds based on the predicted partial point clouds. We explicitly impose the camera model geometrical constraint in our pipeline and enforce the alignment of the generated point clouds and estimated depth maps. Experimental results for the single image 3D object reconstruction task show that the proposed method outperforms existing state-of-the-art methods. Both the qualitative and quantitative results demonstrate the generality and suitability of our method.

Anil S. Baslamisli, Thomas T. Groenestege, Partha Das et al.

Semantic segmentation of outdoor scenes is problematic when there are variations in imaging conditions. It is known that albedo (reflectance) is invariant to all kinds of illumination effects. Thus, using reflectance images for semantic segmentation task can be favorable. Additionally, not only segmentation may benefit from reflectance, but also segmentation may be useful for reflectance computation. Therefore, in this paper, the tasks of semantic segmentation and intrinsic image decomposition are considered as a combined process by exploring their mutual relationship in a joint fashion. To that end, we propose a supervised end-to-end CNN architecture to jointly learn intrinsic image decomposition and semantic segmentation. We analyze the gains of addressing those two problems jointly. Moreover, new cascade CNN architectures for intrinsic-for-segmentation and segmentation-for-intrinsic are proposed as single tasks. Furthermore, a dataset of 35K synthetic images of natural environments is created with corresponding albedo and shading (intrinsics), as well as semantic labels (segmentation) assigned to each object/scene. The experiments show that joint learning of intrinsic image decomposition and semantic segmentation is beneficial for both tasks for natural scenes. Dataset and models are available at: https://ivi.fnwi.uva.nl/cv/intrinseg

Hoang-An Le, Anil S. Baslamisli, Thomas Mensink et al.

Optical flow, semantic segmentation, and surface normals represent different information modalities, yet together they bring better cues for scene understanding problems. In this paper, we study the influence between the three modalities: how one impacts on the others and their efficiency in combination. We employ a modular approach using a convolutional refinement network which is trained supervised but isolated from RGB images to enforce joint modality features. To assist the training process, we create a large-scale synthetic outdoor dataset that supports dense annotation of semantic segmentation, optical flow, and surface normals. The experimental results show positive influence among the three modalities, especially for objects' boundaries, region consistency, and scene structures.