Liangliang Nan

Zhaiyu Chen, Yilei Shi, Liangliang Nan et al.

We present PolyGNN, a polyhedron-based graph neural network for 3D building reconstruction from point clouds. PolyGNN learns to assemble primitives obtained by polyhedral decomposition via graph node classification, achieving a watertight and compact reconstruction. To effectively represent arbitrary-shaped polyhedra in the neural network, we propose a skeleton-based sampling strategy to generate polyhedron-wise queries. These queries are then incorporated with inter-polyhedron adjacency to enhance the classification. PolyGNN is end-to-end optimizable and is designed to accommodate variable-size input points, polyhedra, and queries with an index-driven batching technique. To address the abstraction gap between existing city-building models and the underlying instances, and provide a fair evaluation of the proposed method, we develop our method on a large-scale synthetic dataset with well-defined ground truths of polyhedral labels. We further conduct a transferability analysis across cities and on real-world point clouds. Both qualitative and quantitative results demonstrate the effectiveness of our method, particularly its efficiency for large-scale reconstructions. The source code and data are available at https://github.com/chenzhaiyu/polygnn.

Shenglan Du, Nail Ibrahimli, Jantien Stoter et al.

Feedforward fully convolutional neural networks currently dominate in semantic segmentation of 3D point clouds. Despite their great success, they suffer from the loss of local information at low-level layers, posing significant challenges to accurate scene segmentation and precise object boundary delineation. Prior works either address this issue by post-processing or jointly learn object boundaries to implicitly improve feature encoding of the networks. These approaches often require additional modules which are difficult to integrate into the original architecture. To improve the segmentation near object boundaries, we propose a boundary-aware feature propagation mechanism. This mechanism is achieved by exploiting a multi-task learning framework that aims to explicitly guide the boundaries to their original locations. With one shared encoder, our network outputs (i) boundary localization, (ii) prediction of directions pointing to the object's interior, and (iii) semantic segmentation, in three parallel streams. The predicted boundaries and directions are fused to propagate the learned features to refine the segmentation. We conduct extensive experiments on the S3DIS and SensatUrban datasets against various baseline methods, demonstrating that our proposed approach yields consistent improvements by reducing boundary errors. Our code is available at https://github.com/shenglandu/PushBoundary.

Shiming Wang, Holger Caesar, Liangliang Nan et al.

Multi-sensor object detection is an active research topic in automated driving, but the robustness of such detection models against missing sensor input (modality missing), e.g., due to a sudden sensor failure, is a critical problem which remains under-studied. In this work, we propose UniBEV, an end-to-end multi-modal 3D object detection framework designed for robustness against missing modalities: UniBEV can operate on LiDAR plus camera input, but also on LiDAR-only or camera-only input without retraining. To facilitate its detector head to handle different input combinations, UniBEV aims to create well-aligned Bird's Eye View (BEV) feature maps from each available modality. Unlike prior BEV-based multi-modal detection methods, all sensor modalities follow a uniform approach to resample features from the native sensor coordinate systems to the BEV features. We furthermore investigate the robustness of various fusion strategies w.r.t. missing modalities: the commonly used feature concatenation, but also channel-wise averaging, and a generalization to weighted averaging termed Channel Normalized Weights. To validate its effectiveness, we compare UniBEV to state-of-the-art BEVFusion and MetaBEV on nuScenes over all sensor input combinations. In this setting, UniBEV achieves $52.5 \%$ mAP on average over all input combinations, significantly improving over the baselines ($43.5 \%$ mAP on average for BEVFusion, $48.7 \%$ mAP on average for MetaBEV). An ablation study shows the robustness benefits of fusing by weighted averaging over regular concatenation, and of sharing queries between the BEV encoders of each modality. Our code is available at https://github.com/tudelft-iv/UniBEV.

Zhe Zhu, Liangliang Nan, Haoran Xie et al.

How will you repair a physical object with some missings? You may imagine its original shape from previously captured images, recover its overall (global) but coarse shape first, and then refine its local details. We are motivated to imitate the physical repair procedure to address point cloud completion. To this end, we propose a cross-modal shape-transfer dual-refinement network (termed CSDN), a coarse-to-fine paradigm with images of full-cycle participation, for quality point cloud completion. CSDN mainly consists of "shape fusion" and "dual-refinement" modules to tackle the cross-modal challenge. The first module transfers the intrinsic shape characteristics from single images to guide the geometry generation of the missing regions of point clouds, in which we propose IPAdaIN to embed the global features of both the image and the partial point cloud into completion. The second module refines the coarse output by adjusting the positions of the generated points, where the local refinement unit exploits the geometric relation between the novel and the input points by graph convolution, and the global constraint unit utilizes the input image to fine-tune the generated offset. Different from most existing approaches, CSDN not only explores the complementary information from images but also effectively exploits cross-modal data in the whole coarse-to-fine completion procedure. Experimental results indicate that CSDN performs favorably against ten competitors on the cross-modal benchmark.

Nail Ibrahimli, Hugo Ledoux, Julian Kooij et al.

Traditional MVS methods have good accuracy but struggle with completeness, while recently developed learning-based multi-view stereo (MVS) techniques have improved completeness except accuracy being compromised. We propose depth discontinuity learning for MVS methods, which further improves accuracy while retaining the completeness of the reconstruction. Our idea is to jointly estimate the depth and boundary maps where the boundary maps are explicitly used for further refinement of the depth maps. We validate our idea and demonstrate that our strategies can be easily integrated into the existing learning-based MVS pipeline where the reconstruction depends on high-quality depth map estimation. Extensive experiments on various datasets show that our method improves reconstruction quality compared to baseline. Experiments also demonstrate that the presented model and strategies have good generalization capabilities. The source code will be available soon.

Pei An, Junfeng Ding, Jiaqi Yang et al.

Image-to-point-cloud (I2P) registration aims to align 2D images with 3D point clouds by establishing reliable 2D-3D correspondences. The drastic modality gap between images and point clouds makes it challenging to learn features that are both discriminative and generalizable, leading to severe performance drops in unseen scenarios. We address this challenge by introducing a heterogeneous graph that enables refining both cross-modal features and correspondences within a unified architecture. The proposed graph represents a mapping between segmented 2D and 3D regions, which enhances cross-modal feature interaction and thus improves feature discriminability. In addition, modeling the consistency among vertices and edges within the graph enables pruning of unreliable correspondences. Building on these insights, we propose a heterogeneous graph embedded I2P registration method, termed Hg-I2P. It learns a heterogeneous graph by mining multi-path feature relationships, adapts features under the guidance of heterogeneous edges, and prunes correspondences using graph-based projection consistency. Experiments on six indoor and outdoor benchmarks under cross-domain setups demonstrate that Hg-I2P significantly outperforms existing methods in both generalization and accuracy. Code is released on https://github.com/anpei96/hg-i2p-demo.

Mubariz Zaffar, Liangliang Nan, Julian Francisco Pieter Kooij

Visual Place Recognition (VPR) is an image-based localization method that estimates the camera location of a query image by retrieving the most similar reference image from a map of geo-tagged reference images. In this work, we look into two fundamental bottlenecks for its localization accuracy: reference map sparseness and viewpoint invariance. Firstly, the reference images for VPR are only available at sparse poses in a map, which enforces an upper bound on the maximum achievable localization accuracy through VPR. We therefore propose Continuous Place-descriptor Regression (CoPR) to densify the map and improve localization accuracy. We study various interpolation and extrapolation models to regress additional VPR feature descriptors from only the existing references. Secondly, we compare different feature encoders and show that CoPR presents value for all of them. We evaluate our models on three existing public datasets and report on average around 30% improvement in VPR-based localization accuracy using CoPR, on top of the 15% increase by using a viewpoint-variant loss for the feature encoder. The complementary relation between CoPR and Relative Pose Estimation is also discussed.

Baian Chen, Liangliang Nan, Haoran Xie et al.

Capturing both local and global features of irregular point clouds is essential to 3D object detection (3OD). However, mainstream 3D detectors, e.g., VoteNet and its variants, either abandon considerable local features during pooling operations or ignore many global features in the whole scene context. This paper explores new modules to simultaneously learn local-global features of scene point clouds that serve 3OD positively. To this end, we propose an effective 3OD network via simultaneous local-global feature learning (dubbed 3DLG-Detector). 3DLG-Detector has two key contributions. First, it develops a Dynamic Points Interaction (DPI) module that preserves effective local features during pooling. Besides, DPI is detachable and can be incorporated into existing 3OD networks to boost their performance. Second, it develops a Global Context Aggregation module to aggregate multi-scale features from different layers of the encoder to achieve scene context-awareness. Our method shows improvements over thirteen competitors in terms of detection accuracy and robustness on both the SUN RGB-D and ScanNet datasets. Source code will be available upon publication.

Yancong Lin, Shiming Wang, Liangliang Nan et al.

Scene flow estimation aims to recover per-point motion from two adjacent LiDAR scans. However, in real-world applications such as autonomous driving, points rarely move independently of others, especially for nearby points belonging to the same object, which often share the same motion. Incorporating this locally rigid motion constraint has been a key challenge in self-supervised scene flow estimation, which is often addressed by post-processing or appending extra regularization. While these approaches are able to improve the rigidity of predicted flows, they lack an architectural inductive bias for local rigidity within the model structure, leading to suboptimal learning efficiency and inferior performance. In contrast, we enforce local rigidity with a lightweight add-on module in neural network design, enabling end-to-end learning. We design a discretized voting space that accommodates all possible translations and then identify the one shared by nearby points by differentiable voting. Additionally, to ensure computational efficiency, we operate on pillars rather than points and learn representative features for voting per pillar. We plug the Voting Module into popular model designs and evaluate its benefit on Argoverse 2 and Waymo datasets. We outperform baseline works with only marginal compute overhead. Code is available at https://github.com/tudelft-iv/VoteFlow.

Yun Liu, Peng Li, Xuefeng Yan et al.

The core of self-supervised point cloud learning lies in setting up appropriate pretext tasks, to construct a pre-training framework that enables the encoder to perceive 3D objects effectively. In this paper, we integrate two prevalent methods, masked point modeling (MPM) and 3D-to-2D generation, as pretext tasks within a pre-training framework. We leverage the spatial awareness and precise supervision offered by these two methods to address their respective limitations: ambiguous supervision signals and insensitivity to geometric information. Specifically, the proposed framework, abbreviated as PointCG, consists of a Hidden Point Completion (HPC) module and an Arbitrary-view Image Generation (AIG) module. We first capture visible points from arbitrary views as inputs by removing hidden points. Then, HPC extracts representations of the inputs with an encoder and completes the entire shape with a decoder, while AIG is used to generate rendered images based on the visible points' representations. Extensive experiments demonstrate the superiority of the proposed method over the baselines in various downstream tasks. Our code will be made available upon acceptance.

Xing He, Zhe Zhu, Liangliang Nan et al.

Conventional methods for point cloud completion, typically trained on synthetic datasets, face significant challenges when applied to out-of-distribution real-world scans. In this paper, we propose an effective yet simple source-free domain adaptation framework for point cloud completion, termed \textbf{PointSFDA}. Unlike unsupervised domain adaptation that reduces the domain gap by directly leveraging labeled source data, PointSFDA uses only a pretrained source model and unlabeled target data for adaptation, avoiding the need for inaccessible source data in practical scenarios. Being the first source-free domain adaptation architecture for point cloud completion, our method offers two core contributions. First, we introduce a coarse-to-fine distillation solution to explicitly transfer the global geometry knowledge learned from the source dataset. Second, as noise may be introduced due to domain gaps, we propose a self-supervised partial-mask consistency training strategy to learn local geometry information in the target domain. Extensive experiments have validated that our method significantly improves the performance of state-of-the-art networks in cross-domain shape completion. Our code is available at \emph{\textcolor{magenta}{https://github.com/Starak-x/PointSFDA}}.

Weixiao Gao, Liangliang Nan, Bas Boom et al.

We introduce a novel deep learning-based framework to interpret 3D urban scenes represented as textured meshes. Based on the observation that object boundaries typically align with the boundaries of planar regions, our framework achieves semantic segmentation in two steps: planarity-sensible over-segmentation followed by semantic classification. The over-segmentation step generates an initial set of mesh segments that capture the planar and non-planar regions of urban scenes. In the subsequent classification step, we construct a graph that encodes the geometric and photometric features of the segments in its nodes and the multi-scale contextual features in its edges. The final semantic segmentation is obtained by classifying the segments using a graph convolutional network. Experiments and comparisons on two semantic urban mesh benchmarks demonstrate that our approach outperforms the state-of-the-art methods in terms of boundary quality, mean IoU (intersection over union), and generalization ability. We also introduce several new metrics for evaluating mesh over-segmentation methods dedicated to semantic segmentation, and our proposed over-segmentation approach outperforms state-of-the-art methods on all metrics. Our source code is available at \url{https://github.com/WeixiaoGao/PSSNet}.

Zhaiyu Chen, Hugo Ledoux, Seyran Khademi et al.

While three-dimensional (3D) building models play an increasingly pivotal role in many real-world applications, obtaining a compact representation of buildings remains an open problem. In this paper, we present a novel framework for reconstructing compact, watertight, polygonal building models from point clouds. Our framework comprises three components: (a) a cell complex is generated via adaptive space partitioning that provides a polyhedral embedding as the candidate set; (b) an implicit field is learned by a deep neural network that facilitates building occupancy estimation; (c) a Markov random field is formulated to extract the outer surface of a building via combinatorial optimization. We evaluate and compare our method with state-of-the-art methods in generic reconstruction, model-based reconstruction, geometry simplification, and primitive assembly. Experiments on both synthetic and real-world point clouds have demonstrated that, with our neural-guided strategy, high-quality building models can be obtained with significant advantages in fidelity, compactness, and computational efficiency. Our method also shows robustness to noise and insufficient measurements, and it can directly generalize from synthetic scans to real-world measurements. The source code of this work is freely available at https://github.com/chenzhaiyu/points2poly.

Xufei Wang, Zexin Yang, Xiaojun Cheng et al.

Registering point clouds of forest environments is an essential prerequisite for LiDAR applications in precision forestry. State-of-the-art methods for forest point cloud registration require the extraction of individual tree attributes, and they have an efficiency bottleneck when dealing with point clouds of real-world forests with dense trees. We propose an automatic, robust, and efficient method for the registration of forest point clouds. Our approach first locates tree stems from raw point clouds and then matches the stems based on their relative spatial relationship to determine the registration transformation. The algorithm requires no extra individual tree attributes and has quadratic complexity to the number of trees in the environment, allowing it to align point clouds of large forest environments. Extensive experiments on forest terrestrial point clouds have revealed that our method inherits the effectiveness and robustness of the stem-based registration strategy while exceedingly increasing its efficiency. Besides, we introduce a new benchmark dataset that complements the very few existing open datasets for the development and evaluation of registration methods for forest point clouds. The source code of our method and the dataset are available at https://github.com/zexinyang/GlobalMatch.

Weixiao Gao, Liangliang Nan, Bas Boom et al.

Recent developments in data acquisition technology allow us to collect 3D texture meshes quickly. Those can help us understand and analyse the urban environment, and as a consequence are useful for several applications like spatial analysis and urban planning. Semantic segmentation of texture meshes through deep learning methods can enhance this understanding, but it requires a lot of labelled data. The contributions of this work are threefold: (1) a new benchmark dataset of semantic urban meshes, (2) a novel semi-automatic annotation framework, and (3) an annotation tool for 3D meshes. In particular, our dataset covers about 4 km2 in Helsinki (Finland), with six classes, and we estimate that we save about 600 hours of labelling work using our annotation framework, which includes initial segmentation and interactive refinement. We also compare the performance of several state-of-theart 3D semantic segmentation methods on the new benchmark dataset. Other researchers can use our results to train their networks: the dataset is publicly available, and the annotation tool is released as open-source.

Xueyang Kang, Zijian Yu, Kourosh Khoshelham et al.

Interactive segmentation allows efficient label generation by leveraging user-provided clicks to progressively refine predictions, which is critical when fully supervised labels are costly or generalization to unseen classes is needed. Existing 3D interactive methods are limited: most operate sequentially, predicting only one object per iteration with binary masks, while several recent approaches depend on 2D foundation models and camera alignment to bridge the 2D-3D gap. To address these limitations, we propose a novel interactive segmentation framework that operates directly on sparse, randomly downsampled 3D points and processes multiple object clicks in a single forward pass. Our framework consists of a point Transformer-based encoder and a hierarchical mask decoder, which integrates multi-level crop-and-merge operations conditioned on learnable semantic embeddings. Unlike prior interactive approaches that require repeated model updates after each manually corrective click, our method jointly reasons over all click queries, modeling inter-instance relationships and refining both spatial masks and semantic predictions through spatial and semantic embeddings. Extensive experiments demonstrate that our model improves the mIoU metric by over 20 percent compared to strong baselines and achieves 8-10 percent gains under cross-dataset evaluation for a one-click per instance setting, often requiring only a single click per object. Our approach provides a generalizable and efficient solution for interactive 3D instance segmentation, particularly suitable for real-time applications such as robotic manipulation, navigation, and rapid 3D semantic annotation.

Tomislav Medic, Liangliang Nan

3D instance segmentation for laser scanning (LiDAR) point clouds remains a challenge in many remote sensing-related domains. Successful solutions typically rely on supervised deep learning and manual annotations, and consequently focus on objects that can be well delineated through visual inspection and manual labeling of point clouds. However, for tasks with more complex and cluttered scenes, such as in-field plant phenotyping in agriculture, such approaches are often infeasible. In this study, we tackle the task of in-field wheat head instance segmentation directly from terrestrial laser scanning (TLS) point clouds. To address the problem and circumvent the need for manual annotations, we propose a novel two-stage pipeline. To obtain the initial 3D instance proposals, the first stage uses 3D-to-2D multi-view projections, the Grounded SAM pipeline for zero-shot 2D object-centric segmentation, and multi-view label fusion. The second stage uses these initial proposals as noisy pseudo-labels to train a supervised 3D panoptic-style segmentation neural network. Our results demonstrate the feasibility of the proposed approach and show performance improvementsrelative to Wheat3DGS, a recent alternative solution for in-field wheat head instance segmentation without manual 3D annotations based on multi-view RGB images and 3D Gaussian Splatting, showcasing TLS as a competitive sensing alternative. Moreover, the results show that both stages of the proposed pipeline can deliver usable 3D instance segmentation without manual annotations, indicating promising, low-effort transferability to other comparable TLS-based point cloud segmentation tasks.

Mubariz Zaffar, Liangliang Nan, Julian F. P. Kooij

In Visual Place Recognition (VPR) the pose of a query image is estimated by comparing the image to a map of reference images with known reference poses. As is typical for image retrieval problems, a feature extractor maps the query and reference images to a feature space, where a nearest neighbor search is then performed. However, till recently little attention has been given to quantifying the confidence that a retrieved reference image is a correct match. Highly certain but incorrect retrieval can lead to catastrophic failure of VPR-based localization pipelines. This work compares for the first time the main approaches for estimating the image-matching uncertainty, including the traditional retrieval-based uncertainty estimation, more recent data-driven aleatoric uncertainty estimation, and the compute-intensive geometric verification. We further formulate a simple baseline method, ``SUE'', which unlike the other methods considers the freely-available poses of the reference images in the map. Our experiments reveal that a simple L2-distance between the query and reference descriptors is already a better estimate of image-matching uncertainty than current data-driven approaches. SUE outperforms the other efficient uncertainty estimation methods, and its uncertainty estimates complement the computationally expensive geometric verification approach. Future works for uncertainty estimation in VPR should consider the baselines discussed in this work.

Lipeng Gu, Xuefeng Yan, Liangliang Nan et al.

Current methodologies in point cloud analysis predominantly explore 3D geometries, often achieved through the introduction of intricate learnable geometric extractors in the encoder or by deepening networks with repeated blocks. However, these approaches inevitably lead to a significant number of learnable parameters, resulting in substantial computational costs and imposing memory burdens on CPU/GPU. Additionally, the existing strategies are primarily tailored for object-level point cloud classification and segmentation tasks, with limited extensions to crucial scene-level applications, such as autonomous driving. In response to these limitations, we introduce PointeNet, an efficient network designed specifically for point cloud analysis. PointeNet distinguishes itself with its lightweight architecture, low training cost, and plug-and-play capability, effectively capturing representative features. The network consists of a Multivariate Geometric Encoding (MGE) module and an optional Distance-aware Semantic Enhancement (DSE) module. The MGE module employs operations of sampling, grouping, and multivariate geometric aggregation to lightweightly capture and adaptively aggregate multivariate geometric features, providing a comprehensive depiction of 3D geometries. The DSE module, designed for real-world autonomous driving scenarios, enhances the semantic perception of point clouds, particularly for distant points. Our method demonstrates flexibility by seamlessly integrating with a classification/segmentation head or embedding into off-the-shelf 3D object detection networks, achieving notable performance improvements at a minimal cost. Extensive experiments on object-level datasets, including ModelNet40, ScanObjectNN, ShapeNetPart, and the scene-level dataset KITTI, demonstrate the superior performance of PointeNet over state-of-the-art methods in point cloud analysis.

Nail Ibrahimli, Julian F. P. Kooij, Liangliang Nan

We introduce MuVieCAST, a modular multi-view consistent style transfer network architecture that enables consistent style transfer between multiple viewpoints of the same scene. This network architecture supports both sparse and dense views, making it versatile enough to handle a wide range of multi-view image datasets. The approach consists of three modules that perform specific tasks related to style transfer, namely content preservation, image transformation, and multi-view consistency enforcement. We extensively evaluate our approach across multiple application domains including depth-map-based point cloud fusion, mesh reconstruction, and novel-view synthesis. Our experiments reveal that the proposed framework achieves an exceptional generation of stylized images, exhibiting consistent outcomes across perspectives. A user study focusing on novel-view synthesis further confirms these results, with approximately 68\% of cases participants expressing a preference for our generated outputs compared to the recent state-of-the-art method. Our modular framework is extensible and can easily be integrated with various backbone architectures, making it a flexible solution for multi-view style transfer. More results are demonstrated on our project page: muviecast.github.io

Weixiao Gao, Liangliang Nan, Hugo Ledoux

Semantic segmentation in urban scene analysis has mainly focused on images or point clouds, while textured meshes - offering richer spatial representation - remain underexplored. This paper introduces SUM Parts, the first large-scale dataset for urban textured meshes with part-level semantic labels, covering about 2.5 km2 with 21 classes. The dataset was created using our own annotation tool, which supports both face- and texture-based annotations with efficient interactive selection. We also provide a comprehensive evaluation of 3D semantic segmentation and interactive annotation methods on this dataset. Our project page is available at https://tudelft3d.github.io/SUMParts/.

Xueyang Kang, Zizhao Li, Tian Lan et al.

3D shape anomaly detection is a crucial task for industrial inspection and geometric analysis. Existing deep learning approaches typically learn representations of normal shapes and identify anomalies via out-of-distribution feature detection or decoder-based reconstruction. They often fail to generalize across diverse anomaly types and scales, such as global geometric errors (e.g., planar shifts, angle misalignments), and are sensitive to noisy or incomplete local points during training. To address these limitations, we propose a hierarchical point-patch anomaly scoring network that jointly models regional part features and local point features for robust anomaly reasoning. An adaptive patchification module integrates self-supervised decomposition to capture complex structural deviations. Beyond evaluations on public benchmarks (Anomaly-ShapeNet and Real3D-AD), we release an industrial test set with real CAD models exhibiting planar, angular, and structural defects. Experiments on public and industrial datasets show superior AUC-ROC and AUC-PR performance, including over 40% point-level improvement on the new industrial anomaly type and average object-level gains of 7% on Real3D-AD and 4% on Anomaly-ShapeNet, demonstrating strong robustness and generalization.

Zhaiyu Chen, Yuqing Wang, Liangliang Nan et al.

Existing polygonal surface reconstruction methods heavily depend on input completeness and struggle with incomplete point clouds. We argue that while current point cloud completion techniques may recover missing points, they are not optimized for polygonal surface reconstruction, where the parametric representation of underlying surfaces remains overlooked. To address this gap, we introduce parametric completion, a novel paradigm for point cloud completion, which recovers parametric primitives instead of individual points to convey high-level geometric structures. Our presented approach, PaCo, enables high-quality polygonal surface reconstruction by leveraging plane proxies that encapsulate both plane parameters and inlier points, proving particularly effective in challenging scenarios with highly incomplete data. Comprehensive evaluations of our approach on the ABC dataset establish its effectiveness with superior performance and set a new standard for polygonal surface reconstruction from incomplete data. Project page: https://parametric-completion.github.io.

Lipeng Gu, Xuefeng Yan, Weiming Wang et al.

The fusion of camera- and LiDAR-based detections offers a promising solution to mitigate tracking failures in 3D multi-object tracking (MOT). However, existing methods predominantly exploit camera detections to correct tracking failures caused by potential LiDAR detection problems, neglecting the reciprocal benefit of refining camera detections using LiDAR data. This limitation is rooted in their single-stage architecture, akin to single-stage object detectors, lacking a dedicated trajectory refinement module to fully exploit the complementary multi-modal information. To this end, we introduce CrossTracker, a novel two-stage paradigm for online multi-modal 3D MOT. CrossTracker operates in a coarse-to-fine manner, initially generating coarse trajectories and subsequently refining them through an independent refinement process. Specifically, CrossTracker incorporates three essential modules: i) a multi-modal modeling (M^3) module that, by fusing multi-modal information (images, point clouds, and even plane geometry extracted from images), provides a robust metric for subsequent trajectory generation. ii) a coarse trajectory generation (C-TG) module that generates initial coarse dual-stream trajectories, and iii) a trajectory refinement (TR) module that refines coarse trajectories through cross correction between camera and LiDAR streams. Comprehensive experiments demonstrate the superior performance of our CrossTracker over its eighteen competitors, underscoring its effectiveness in harnessing the synergistic benefits of camera and LiDAR sensors for robust multi-modal 3D MOT.

Guangpeng Fan, Fei Yan, Xiangquan Zeng et al.

We have developed the world's first canopy height map of the distribution area of world-level giant trees. This mapping is crucial for discovering more individual and community world-level giant trees, and for analyzing and quantifying the effectiveness of biodiversity conservation measures in the Yarlung Tsangpo Grand Canyon (YTGC) National Nature Reserve. We proposed a method to map the canopy height of the primeval forest within the world-level giant tree distribution area by using a spaceborne LiDAR fusion satellite imagery (Global Ecosystem Dynamics Investigation (GEDI), ICESat-2, and Sentinel-2) driven deep learning modeling. And we customized a pyramid receptive fields depth separable CNN (PRFXception). PRFXception, a CNN architecture specifically customized for mapping primeval forest canopy height to infer the canopy height at the footprint level of GEDI and ICESat-2 from Sentinel-2 optical imagery with a 10-meter spatial resolution. We conducted a field survey of 227 permanent plots using a stratified sampling method and measured several giant trees using UAV-LS. The predicted canopy height was compared with ICESat-2 and GEDI validation data (RMSE =7.56 m, MAE=6.07 m, ME=-0.98 m, R^2=0.58 m), UAV-LS point clouds (RMSE =5.75 m, MAE =3.72 m, ME = 0.82 m, R^2= 0.65 m), and ground survey data (RMSE = 6.75 m, MAE = 5.56 m, ME= 2.14 m, R^2=0.60 m). We mapped the potential distribution map of world-level giant trees and discovered two previously undetected giant tree communities with an 89% probability of having trees 80-100 m tall, potentially taller than Asia's tallest tree. This paper provides scientific evidence confirming southeastern Tibet--northwestern Yunnan as the fourth global distribution center of world-level giant trees initiatives and promoting the inclusion of the YTGC giant tree distribution area within the scope of China's national park conservation.

Xin Li, Peng Li, Zeyong Wei et al.

Introducing BERT into cross-modal settings raises difficulties in its optimization for handling multiple modalities. Both the BERT architecture and training objective need to be adapted to incorporate and model information from different modalities. In this paper, we address these challenges by exploring the implicit semantic and geometric correlations between 2D and 3D data of the same objects/scenes. We propose a new cross-modal BERT-style self-supervised learning paradigm, called Cross-BERT. To facilitate pretraining for irregular and sparse point clouds, we design two self-supervised tasks to boost cross-modal interaction. The first task, referred to as Point-Image Alignment, aims to align features between unimodal and cross-modal representations to capture the correspondences between the 2D and 3D modalities. The second task, termed Masked Cross-modal Modeling, further improves mask modeling of BERT by incorporating high-dimensional semantic information obtained by cross-modal interaction. By performing cross-modal interaction, Cross-BERT can smoothly reconstruct the masked tokens during pretraining, leading to notable performance enhancements for downstream tasks. Through empirical evaluation, we demonstrate that Cross-BERT outperforms existing state-of-the-art methods in 3D downstream applications. Our work highlights the effectiveness of leveraging cross-modal 2D knowledge to strengthen 3D point cloud representation and the transferable capability of BERT across modalities.

Shiming Wang, Holger Caesar, Liangliang Nan et al.

In autonomous driving, multi-modal perception tasks like 3D object detection typically rely on well-synchronized sensors, both at training and inference. However, despite the use of hardware- or software-based synchronization algorithms, perfect synchrony is rarely guaranteed: Sensors may operate at different frequencies, and real-world factors such as network latency, hardware failures, or processing bottlenecks often introduce time offsets between sensors. Such asynchrony degrades perception performance, especially for dynamic objects. To address this challenge, we propose AsyncBEV, a trainable lightweight and generic module to improve the robustness of 3D Birds' Eye View (BEV) object detection models against sensor asynchrony. Inspired by scene flow estimation, AsyncBEV first estimates the 2D flow from the BEV features of two different sensor modalities, taking into account the known time offset between these sensor measurements. The predicted feature flow is then used to warp and spatially align the feature maps, which we show can easily be integrated into different current BEV detector architectures (e.g., BEV grid-based and token-based). Extensive experiments demonstrate AsyncBEV improves robustness against both small and large asynchrony between LiDAR or camera sensors in both the token-based CMT and grid-based UniBEV, especially for dynamic objects. We significantly outperform the ego motion compensated CMT and UniBEV baselines, notably by $16.6$ % and $11.9$ % NDS on dynamic objects in the worst-case scenario of a $0.5 s$ time offset. Code will be released upon acceptance.

Mubariz Zaffar, Liangliang Nan, Sebastian Scherer et al.

Given a query image, Visual Place Recognition (VPR) is the task of retrieving an image of the same place from a reference database with robustness to viewpoint and appearance changes. Recent works show that some VPR benchmarks are solved by methods using Vision-Foundation-Model backbones and trained on large-scale and diverse VPR-specific datasets. Several benchmarks remain challenging, particularly when the test environments differ significantly from the usual VPR training datasets. We propose a complementary, unexplored source of information to bridge the train-test domain gap, which can further improve the performance of State-of-the-Art (SOTA) VPR methods on such challenging benchmarks. Concretely, we identify that the test-time reference set, the "map", contains images and poses of the target domain, and must be available before the test-time query is received in several VPR applications. Therefore, we propose to perform simple Reference-Set-Finetuning (RSF) of VPR models on the map, boosting the SOTA (~2.3% increase on average for Recall@1) on these challenging datasets. Finetuned models retain generalization, and RSF works across diverse test datasets.

Pei An, Jiaqi Yang, Muyao Peng et al.

Image-to-point-cloud (I2P) registration is a fundamental problem in computer vision, focusing on establishing 2D-3D correspondences between an image and a point cloud. The differential perspective-n-point (PnP) has been widely used to supervise I2P registration networks by enforcing the projective constraints on 2D-3D correspondences. However, differential PnP is highly sensitive to noise and outliers in the predicted correspondences. This issue hinders the effectiveness of correspondence learning. Inspired by the robustness of blind PnP against noise and outliers in correspondences, we propose an approximated blind PnP based correspondence learning approach. To mitigate the high computational cost of blind PnP, we simplify blind PnP to an amenable task of minimizing Chamfer distance between learned 2D and 3D keypoints, called MinCD-PnP. To effectively solve MinCD-PnP, we design a lightweight multi-task learning module, named as MinCD-Net, which can be easily integrated into the existing I2P registration architectures. Extensive experiments on 7-Scenes, RGBD-V2, ScanNet, and self-collected datasets demonstrate that MinCD-Net outperforms state-of-the-art methods and achieves a higher inlier ratio (IR) and registration recall (RR) in both cross-scene and cross-dataset settings.

Yu Qi, Lipeng Gu, Honghua Chen et al.

Existing 3D visual grounding methods rely on precise text prompts to locate objects within 3D scenes. Speech, as a natural and intuitive modality, offers a promising alternative. Real-world speech inputs, however, often suffer from transcription errors due to accents, background noise, and varying speech rates, limiting the applicability of existing 3DVG methods. To address these challenges, we propose \textbf{SpeechRefer}, a novel 3DVG framework designed to enhance performance in the presence of noisy and ambiguous speech-to-text transcriptions. SpeechRefer integrates seamlessly with xisting 3DVG models and introduces two key innovations. First, the Speech Complementary Module captures acoustic similarities between phonetically related words and highlights subtle distinctions, generating complementary proposal scores from the speech signal. This reduces dependence on potentially erroneous transcriptions. Second, the Contrastive Complementary Module employs contrastive learning to align erroneous text features with corresponding speech features, ensuring robust performance even when transcription errors dominate. Extensive experiments on the SpeechRefer and peechNr3D datasets demonstrate that SpeechRefer improves the performance of existing 3DVG methods by a large margin, which highlights SpeechRefer's potential to bridge the gap between noisy speech inputs and reliable 3DVG, enabling more intuitive and practical multimodal systems.

Yinuo Zheng, Lipeng Gu, Honghua Chen et al.

3D visual grounding (3DVG) is a critical task in scene understanding that aims to identify objects in 3D scenes based on text descriptions. However, existing methods rely on separately pre-trained vision and text encoders, resulting in a significant gap between the two modalities in terms of spatial geometry and semantic categories. This discrepancy often causes errors in object positioning and classification. The paper proposes UniSpace-3D, which innovatively introduces a unified representation space for 3DVG, effectively bridging the gap between visual and textual features. Specifically, UniSpace-3D incorporates three innovative designs: i) a unified representation encoder that leverages the pre-trained CLIP model to map visual and textual features into a unified representation space, effectively bridging the gap between the two modalities; ii) a multi-modal contrastive learning module that further reduces the modality gap; iii) a language-guided query selection module that utilizes the positional and semantic information to identify object candidate points aligned with textual descriptions. Extensive experiments demonstrate that UniSpace-3D outperforms baseline models by at least 2.24% on the ScanRefer and Nr3D/Sr3D datasets. The code will be made available upon acceptance of the paper.

Yabin Xu, Liangliang Nan, Laishui Zhou et al.

Reconstruction of high-fidelity 3D objects or scenes is a fundamental research problem. Recent advances in RGB-D fusion have demonstrated the potential of producing 3D models from consumer-level RGB-D cameras. However, due to the discrete nature and limited resolution of their surface representations (e.g., point- or voxel-based), existing approaches suffer from the accumulation of errors in camera tracking and distortion in the reconstruction, which leads to an unsatisfactory 3D reconstruction. In this paper, we present a method using on-the-fly implicits of Hermite Radial Basis Functions (HRBFs) as a continuous surface representation for camera tracking in an existing RGB-D fusion framework. Furthermore, curvature estimation and confidence evaluation are coherently derived from the inherent surface properties of the on-the-fly HRBF implicits, which devote to a data fusion with better quality. We argue that our continuous but on-the-fly surface representation can effectively mitigate the impact of noise with its robustness and constrain the reconstruction with inherent surface smoothness when being compared with discrete representations. Experimental results on various real-world and synthetic datasets demonstrate that our HRBF-fusion outperforms the state-of-the-art approaches in terms of tracking robustness and reconstruction accuracy.

Jin Huang, Jantien Stoter, Ravi Peters et al.

We present a fully automatic approach for reconstructing compact 3D building models from large-scale airborne point clouds. A major challenge of urban reconstruction from airborne LiDAR point clouds lies in that the vertical walls are typically missing. Based on the observation that urban buildings typically consist of planar roofs connected with vertical walls to the ground, we propose an approach to infer the vertical walls directly from the data. With the planar segments of both roofs and walls, we hypothesize the faces of the building surface, and the final model is obtained by using an extended hypothesis-and-selection-based polygonal surface reconstruction framework. Specifically, we introduce a new energy term to encourage roof preferences and two additional hard constraints into the optimization step to ensure correct topology and enhance detail recovery. Experiments on various large-scale airborne LiDAR point clouds have demonstrated that the method is superior to the state-of-the-art methods in terms of reconstruction accuracy and robustness. In addition, we have generated a new dataset with our method consisting of the point clouds and 3D models of 20k real-world buildings. We believe this dataset can stimulate research in urban reconstruction from airborne LiDAR point clouds and the use of 3D city models in urban applications.

Jiazhou Chen, Yanghui Xu, Shufang Lu et al.

We present a novel 3D instance segmentation framework for Multi-View Stereo (MVS) buildings in urban scenes. Unlike existing works focusing on semantic segmentation of urban scenes, the emphasis of this work lies in detecting and segmenting 3D building instances even if they are attached and embedded in a large and imprecise 3D surface model. Multi-view RGB images are first enhanced to RGBH images by adding a heightmap and are segmented to obtain all roof instances using a fine-tuned 2D instance segmentation neural network. Instance masks from different multi-view images are then clustered into global masks. Our mask clustering accounts for spatial occlusion and overlapping, which can eliminate segmentation ambiguities among multi-view images. Based on these global masks, 3D roof instances are segmented out by mask back-projections and extended to the entire building instances through a Markov random field optimization. A new dataset that contains instance-level annotation for both 3D urban scenes (roofs and buildings) and drone images (roofs) is provided. To the best of our knowledge, it is the first outdoor dataset dedicated to 3D instance segmentation with much more annotations of attached 3D buildings than existing datasets. Quantitative evaluations and ablation studies have shown the effectiveness of all major steps and the advantages of our multi-view framework over the orthophoto-based method.

Bingtao Ma, Hongsen Liu, Liangliang Nan et al.

The 3D mesh is an important representation of geometric data. In the generation of mesh data, geometric deficiencies (e.g., duplicate elements, degenerate faces, isolated vertices, self-intersection, and inner faces) are unavoidable and may violate the topology structure of an object. In this paper, we propose an effective and efficient geometric deficiency elimination algorithm for 3D meshes. Specifically, duplicate elements can be eliminated by assessing the occurrence times of vertices or faces; degenerate faces can be removed according to the outer product of two edges; since isolated vertices do not appear in any face vertices, they can be deleted directly; self-intersecting faces are detected using an AABB tree and remeshed afterward; by simulating whether multiple random rays that shoot from a face can reach infinity, we can judge whether the surface is an inner face, then decide to delete it or not. Experiments on ModelNet40 dataset illustrate that our method can eliminate the deficiencies of the 3D mesh thoroughly.