Cheng Wang

Hai Wu, Chenglu Wen, Shaoshuai Shi et al.

Recently, virtual/pseudo-point-based 3D object detection that seamlessly fuses RGB images and LiDAR data by depth completion has gained great attention. However, virtual points generated from an image are very dense, introducing a huge amount of redundant computation during detection. Meanwhile, noises brought by inaccurate depth completion significantly degrade detection precision. This paper proposes a fast yet effective backbone, termed VirConvNet, based on a new operator VirConv (Virtual Sparse Convolution), for virtual-point-based 3D object detection. VirConv consists of two key designs: (1) StVD (Stochastic Voxel Discard) and (2) NRConv (Noise-Resistant Submanifold Convolution). StVD alleviates the computation problem by discarding large amounts of nearby redundant voxels. NRConv tackles the noise problem by encoding voxel features in both 2D image and 3D LiDAR space. By integrating VirConv, we first develop an efficient pipeline VirConv-L based on an early fusion design. Then, we build a high-precision pipeline VirConv-T based on a transformed refinement scheme. Finally, we develop a semi-supervised pipeline VirConv-S based on a pseudo-label framework. On the KITTI car 3D detection test leaderboard, our VirConv-L achieves 85% AP with a fast running speed of 56ms. Our VirConv-T and VirConv-S attains a high-precision of 86.3% and 87.2% AP, and currently rank 2nd and 1st, respectively. The code is available at https://github.com/hailanyi/VirConv.

Qing Li, Yu-Shen Liu, Jin-San Cheng et al. · tsinghua

We propose a novel normal estimation method called HSurf-Net, which can accurately predict normals from point clouds with noise and density variations. Previous methods focus on learning point weights to fit neighborhoods into a geometric surface approximated by a polynomial function with a predefined order, based on which normals are estimated. However, fitting surfaces explicitly from raw point clouds suffers from overfitting or underfitting issues caused by inappropriate polynomial orders and outliers, which significantly limits the performance of existing methods. To address these issues, we introduce hyper surface fitting to implicitly learn hyper surfaces, which are represented by multi-layer perceptron (MLP) layers that take point features as input and output surface patterns in a high dimensional feature space. We introduce a novel space transformation module, which consists of a sequence of local aggregation layers and global shift layers, to learn an optimal feature space, and a relative position encoding module to effectively convert point clouds into the learned feature space. Our model learns hyper surfaces from the noise-less features and directly predicts normal vectors. We jointly optimize the MLP weights and module parameters in a data-driven manner to make the model adaptively find the most suitable surface pattern for various points. Experimental results show that our HSurf-Net achieves the state-of-the-art performance on the synthetic shape dataset, the real-world indoor and outdoor scene datasets. The code, data and pretrained models are publicly available.

Yifan Liu, Wuyang Li, Cheng Wang et al.

Tooth point cloud segmentation is a fundamental task in many orthodontic applications. Current research mainly focuses on fully supervised learning which demands expensive and tedious manual point-wise annotation. Although recent weakly-supervised alternatives are proposed to use weak labels for 3D segmentation and achieve promising results, they tend to fail when the labels are extremely sparse. Inspired by the powerful promptable segmentation capability of the Segment Anything Model (SAM), we propose a framework named SAMTooth that leverages such capacity to complement the extremely sparse supervision. To automatically generate appropriate point prompts for SAM, we propose a novel Confidence-aware Prompt Generation strategy, where coarse category predictions are aggregated with confidence-aware filtering. Furthermore, to fully exploit the structural and shape clues in SAM's outputs for assisting the 3D feature learning, we advance a Mask-guided Representation Learning that re-projects the generated tooth masks of SAM into 3D space and constrains these points of different teeth to possess distinguished representations. To demonstrate the effectiveness of the framework, we conduct experiments on the public dataset and surprisingly find with only 0.1\% annotations (one point per tooth), our method can surpass recent weakly supervised methods by a large margin, and the performance is even comparable to the recent fully-supervised methods, showcasing the significant potential of applying SAM to 3D perception tasks with sparse labels. Code is available at https://github.com/CUHK-AIM-Group/SAMTooth.

Changjie Qiu, Zhiyong Wang, Xiuhong Lin et al.

Existing point cloud modeling datasets primarily express the modeling precision by pose or trajectory precision rather than the point cloud modeling effect itself. Under this demand, we first independently construct a set of LiDAR system with an optical stage, and then we build a HPMB dataset based on the constructed LiDAR system, a High-Precision, Multi-Beam, real-world dataset. Second, we propose an modeling evaluation method based on HPMB for object-level modeling to overcome this limitation. In addition, the existing point cloud modeling methods tend to generate continuous skeletons of the global environment, hence lacking attention to the shape of complex objects. To tackle this challenge, we propose a novel learning-based joint framework, DSMNet, for high-precision 3D surface modeling from sparse point cloud frames. DSMNet comprises density-aware Point Cloud Registration (PCR) and geometry-aware Point Cloud Sampling (PCS) to effectively learn the implicit structure feature of sparse point clouds. Extensive experiments demonstrate that DSMNet outperforms the state-of-the-art methods in PCS and PCR on Multi-View Partial Point Cloud (MVP) database. Furthermore, the experiments on the open source KITTI and our proposed HPMB datasets show that DSMNet can be generalized as a post-processing of Simultaneous Localization And Mapping (SLAM), thereby improving modeling precision in environments with sparse point clouds.

Hai Wu, Chenglu Wen, Wei Li et al.

3D object detection received increasing attention in autonomous driving recently. Objects in 3D scenes are distributed with diverse orientations. Ordinary detectors do not explicitly model the variations of rotation and reflection transformations. Consequently, large networks and extensive data augmentation are required for robust detection. Recent equivariant networks explicitly model the transformation variations by applying shared networks on multiple transformed point clouds, showing great potential in object geometry modeling. However, it is difficult to apply such networks to 3D object detection in autonomous driving due to its large computation cost and slow reasoning speed. In this work, we present TED, an efficient Transformation-Equivariant 3D Detector to overcome the computation cost and speed issues. TED first applies a sparse convolution backbone to extract multi-channel transformation-equivariant voxel features; and then aligns and aggregates these equivariant features into lightweight and compact representations for high-performance 3D object detection. On the highly competitive KITTI 3D car detection leaderboard, TED ranked 1st among all submissions with competitive efficiency.

Yudi Dai, Yitai Lin, Xiping Lin et al.

We present SLOPER4D, a novel scene-aware dataset collected in large urban environments to facilitate the research of global human pose estimation (GHPE) with human-scene interaction in the wild. Employing a head-mounted device integrated with a LiDAR and camera, we record 12 human subjects' activities over 10 diverse urban scenes from an egocentric view. Frame-wise annotations for 2D key points, 3D pose parameters, and global translations are provided, together with reconstructed scene point clouds. To obtain accurate 3D ground truth in such large dynamic scenes, we propose a joint optimization method to fit local SMPL meshes to the scene and fine-tune the camera calibration during dynamic motions frame by frame, resulting in plausible and scene-natural 3D human poses. Eventually, SLOPER4D consists of 15 sequences of human motions, each of which has a trajectory length of more than 200 meters (up to 1,300 meters) and covers an area of more than 2,000 $m^2$ (up to 13,000 $m^2$), including more than 100K LiDAR frames, 300k video frames, and 500K IMU-based motion frames. With SLOPER4D, we provide a detailed and thorough analysis of two critical tasks, including camera-based 3D HPE and LiDAR-based 3D HPE in urban environments, and benchmark a new task, GHPE. The in-depth analysis demonstrates SLOPER4D poses significant challenges to existing methods and produces great research opportunities. The dataset and code are released at \url{http://www.lidarhumanmotion.net/sloper4d/}

Jialian Li, Jingyi Zhang, Zhiyong Wang et al.

Existing motion capture datasets are largely short-range and cannot yet fit the need of long-range applications. We propose LiDARHuman26M, a new human motion capture dataset captured by LiDAR at a much longer range to overcome this limitation. Our dataset also includes the ground truth human motions acquired by the IMU system and the synchronous RGB images. We further present a strong baseline method, LiDARCap, for LiDAR point cloud human motion capture. Specifically, we first utilize PointNet++ to encode features of points and then employ the inverse kinematics solver and SMPL optimizer to regress the pose through aggregating the temporally encoded features hierarchically. Quantitative and qualitative experiments show that our method outperforms the techniques based only on RGB images. Ablation experiments demonstrate that our dataset is challenging and worthy of further research. Finally, the experiments on the KITTI Dataset and the Waymo Open Dataset show that our method can be generalized to different LiDAR sensor settings.

Yudi Dai, Yitai Lin, Chenglu Wen et al.

We propose Human-centered 4D Scene Capture (HSC4D) to accurately and efficiently create a dynamic digital world, containing large-scale indoor-outdoor scenes, diverse human motions, and rich interactions between humans and environments. Using only body-mounted IMUs and LiDAR, HSC4D is space-free without any external devices' constraints and map-free without pre-built maps. Considering that IMUs can capture human poses but always drift for long-period use, while LiDAR is stable for global localization but rough for local positions and orientations, HSC4D makes both sensors complement each other by a joint optimization and achieves promising results for long-term capture. Relationships between humans and environments are also explored to make their interaction more realistic. To facilitate many down-stream tasks, like AR, VR, robots, autonomous driving, etc., we propose a dataset containing three large scenes (1k-5k $m^2$) with accurate dynamic human motions and locations. Diverse scenarios (climbing gym, multi-story building, slope, etc.) and challenging human activities (exercising, walking up/down stairs, climbing, etc.) demonstrate the effectiveness and the generalization ability of HSC4D. The dataset and code are available at http://www.lidarhumanmotion.net/hsc4d/.

Zimeng Li, Sa Xiao, Cheng Wang et al. · amazon-science

Magnetic resonance imaging (MRI) using hyperpolarized noble gases provides a way to visualize the structure and function of human lung, but the long imaging time limits its broad research and clinical applications. Deep learning has demonstrated great potential for accelerating MRI by reconstructing images from undersampled data. However, most existing deep conventional neural networks (CNN) directly apply square convolution to k-space data without considering the inherent properties of k-space sampling, limiting k-space learning efficiency and image reconstruction quality. In this work, we propose an encoding enhanced (EN2) complex CNN for highly undersampled pulmonary MRI reconstruction. EN2 employs convolution along either the frequency or phase-encoding direction, resembling the mechanisms of k-space sampling, to maximize the utilization of the encoding correlation and integrity within a row or column of k-space. We also employ complex convolution to learn rich representations from the complex k-space data. In addition, we develop a feature-strengthened modularized unit to further boost the reconstruction performance. Experiments demonstrate that our approach can accurately reconstruct hyperpolarized 129Xe and 1H lung MRI from 6-fold undersampled k-space data and provide lung function measurements with minimal biases compared with fully-sampled image. These results demonstrate the effectiveness of the proposed algorithmic components and indicate that the proposed approach could be used for accelerated pulmonary MRI in research and clinical lung disease patient care.

Ming Yan, Xin Wang, Yudi Dai et al.

Motion capture is a long-standing research problem. Although it has been studied for decades, the majority of research focus on ground-based movements such as walking, sitting, dancing, etc. Off-grounded actions such as climbing are largely overlooked. As an important type of action in sports and firefighting field, the climbing movements is challenging to capture because of its complex back poses, intricate human-scene interactions, and difficult global localization. The research community does not have an in-depth understanding of the climbing action due to the lack of specific datasets. To address this limitation, we collect CIMI4D, a large rock \textbf{C}l\textbf{I}mbing \textbf{M}ot\textbf{I}on dataset from 12 persons climbing 13 different climbing walls. The dataset consists of around 180,000 frames of pose inertial measurements, LiDAR point clouds, RGB videos, high-precision static point cloud scenes, and reconstructed scene meshes. Moreover, we frame-wise annotate touch rock holds to facilitate a detailed exploration of human-scene interaction. The core of this dataset is a blending optimization process, which corrects for the pose as it drifts and is affected by the magnetic conditions. To evaluate the merit of CIMI4D, we perform four tasks which include human pose estimations (with/without scene constraints), pose prediction, and pose generation. The experimental results demonstrate that CIMI4D presents great challenges to existing methods and enables extensive research opportunities. We share the dataset with the research community in http://www.lidarhumanmotion.net/cimi4d/.

Xiao Ma, Guang Zheng, Chi Xu et al.

Characterizing urban environments with broad coverages and high precision is more important than ever for achieving the UN's Sustainable Development Goals (SDGs) as half of the world's populations are living in cities. Urban building height as a fundamental 3D urban structural feature has far-reaching applications. However, so far, producing readily available datasets of recent urban building heights with fine spatial resolutions and global coverages remains a challenging task. Here, we provide an up-to-date global product of urban building heights based on a fine grid size of 150 m around 2020 by combining the spaceborne lidar instrument of GEDI and multi-sourced data including remotely sensed images (i.e., Landsat-8, Sentinel-2, and Sentinel-1) and topographic data. Our results revealed that the estimated method of building height samples based on the GEDI data was effective with 0.78 of Pearson's r and 3.67 m of RMSE in comparison to the reference data. The mapping product also demonstrated good performance as indicated by its strong correlation with the reference data (i.e., Pearson's r = 0.71, RMSE = 4.60 m). Compared with the currently existing products, our global urban building height map holds the ability to provide a higher spatial resolution (i.e., 150 m) with a great level of inherent details about the spatial heterogeneity and flexibility of updating using the GEDI samples as inputs. This work will boost future urban studies across many fields including climate, environmental, ecological, and social sciences.

Yudi Dai, Zhiyong Wang, Xiping Lin et al.

We introduce HiSC4D, a novel Human-centered interaction and 4D Scene Capture method, aimed at accurately and efficiently creating a dynamic digital world, containing large-scale indoor-outdoor scenes, diverse human motions, rich human-human interactions, and human-environment interactions. By utilizing body-mounted IMUs and a head-mounted LiDAR, HiSC4D can capture egocentric human motions in unconstrained space without the need for external devices and pre-built maps. This affords great flexibility and accessibility for human-centered interaction and 4D scene capturing in various environments. Taking into account that IMUs can capture human spatially unrestricted poses but are prone to drifting for long-period using, and while LiDAR is stable for global localization but rough for local positions and orientations, HiSC4D employs a joint optimization method, harmonizing all sensors and utilizing environment cues, yielding promising results for long-term capture in large scenes. To promote research of egocentric human interaction in large scenes and facilitate downstream tasks, we also present a dataset, containing 8 sequences in 4 large scenes (200 to 5,000 $m^2$), providing 36k frames of accurate 4D human motions with SMPL annotations and dynamic scenes, 31k frames of cropped human point clouds, and scene mesh of the environment. A variety of scenarios, such as the basketball gym and commercial street, alongside challenging human motions, such as daily greeting, one-on-one basketball playing, and tour guiding, demonstrate the effectiveness and the generalization ability of HiSC4D. The dataset and code will be publicated on www.lidarhumanmotion.net/hisc4d available for research purposes.

Chenxin Li, Brandon Y. Feng, Yifan Liu et al.

3D reconstruction of biological tissues from a collection of endoscopic images is a key to unlock various important downstream surgical applications with 3D capabilities. Existing methods employ various advanced neural rendering techniques for photorealistic view synthesis, but they often struggle to recover accurate 3D representations when only sparse observations are available, which is usually the case in real-world clinical scenarios. To tackle this {sparsity} challenge, we propose a framework leveraging the prior knowledge from multiple foundation models during the reconstruction process, dubbed as \textit{EndoSparse}. Experimental results indicate that our proposed strategy significantly improves the geometric and appearance quality under challenging sparse-view conditions, including using only three views. In rigorous benchmarking experiments against state-of-the-art methods, \textit{EndoSparse} achieves superior results in terms of accurate geometry, realistic appearance, and rendering efficiency, confirming the robustness to sparse-view limitations in endoscopic reconstruction. \textit{EndoSparse} signifies a steady step towards the practical deployment of neural 3D reconstruction in real-world clinical scenarios. Project page: https://endo-sparse.github.io/.

Hai Wu, Shijia Zhao, Xun Huang et al.

The prevalent approaches of unsupervised 3D object detection follow cluster-based pseudo-label generation and iterative self-training processes. However, the challenge arises due to the sparsity of LiDAR scans, which leads to pseudo-labels with erroneous size and position, resulting in subpar detection performance. To tackle this problem, this paper introduces a Commonsense Prototype-based Detector, termed CPD, for unsupervised 3D object detection. CPD first constructs Commonsense Prototype (CProto) characterized by high-quality bounding box and dense points, based on commonsense intuition. Subsequently, CPD refines the low-quality pseudo-labels by leveraging the size prior from CProto. Furthermore, CPD enhances the detection accuracy of sparsely scanned objects by the geometric knowledge from CProto. CPD outperforms state-of-the-art unsupervised 3D detectors on Waymo Open Dataset (WOD), PandaSet, and KITTI datasets by a large margin. Besides, by training CPD on WOD and testing on KITTI, CPD attains 90.85% and 81.01% 3D Average Precision on easy and moderate car classes, respectively. These achievements position CPD in close proximity to fully supervised detectors, highlighting the significance of our method. The code will be available at https://github.com/hailanyi/CPD.

Zhimin Yuan, Wankang Zeng, Yanfei Su et al.

3D synthetic-to-real unsupervised domain adaptive segmentation is crucial to annotating new domains. Self-training is a competitive approach for this task, but its performance is limited by different sensor sampling patterns (i.e., variations in point density) and incomplete training strategies. In this work, we propose a density-guided translator (DGT), which translates point density between domains, and integrates it into a two-stage self-training pipeline named DGT-ST. First, in contrast to existing works that simultaneously conduct data generation and feature/output alignment within unstable adversarial training, we employ the non-learnable DGT to bridge the domain gap at the input level. Second, to provide a well-initialized model for self-training, we propose a category-level adversarial network in stage one that utilizes the prototype to prevent negative transfer. Finally, by leveraging the designs above, a domain-mixed self-training method with source-aware consistency loss is proposed in stage two to narrow the domain gap further. Experiments on two synthetic-to-real segmentation tasks (SynLiDAR $\rightarrow$ semanticKITTI and SynLiDAR $\rightarrow$ semanticPOSS) demonstrate that DGT-ST outperforms state-of-the-art methods, achieving 9.4$\%$ and 4.3$\%$ mIoU improvements, respectively. Code is available at \url{https://github.com/yuan-zm/DGT-ST}.

Qiming Xia, Wenkai Lin, Haoen Xiang et al.

Unsupervised 3D object detection serves as an important solution for offline 3D object annotation. However, due to the data sparsity and limited views, the clustering-based label fitting in unsupervised object detection often generates low-quality pseudo-labels. Multi-agent collaborative dataset, which involves the sharing of complementary observations among agents, holds the potential to break through this bottleneck. In this paper, we introduce a novel unsupervised method that learns to Detect Objects from Multi-Agent LiDAR scans, termed DOtA, without using labels from external. DOtA first uses the internally shared ego-pose and ego-shape of collaborative agents to initialize the detector, leveraging the generalization performance of neural networks to infer preliminary labels. Subsequently,DOtA uses the complementary observations between agents to perform multi-scale encoding on preliminary labels, then decodes high-quality and low-quality labels. These labels are further used as prompts to guide a correct feature learning process, thereby enhancing the performance of the unsupervised object detection task. Extensive experiments on the V2V4Real and OPV2V datasets show that our DOtA outperforms state-of-the-art unsupervised 3D object detection methods. Additionally, we also validate the effectiveness of the DOtA labels under various collaborative perception frameworks.The code is available at https://github.com/xmuqimingxia/DOtA.

Chuanpan Zheng, Xiaoliang Fan, Cheng Wang et al.

Long-term traffic prediction is highly challenging due to the complexity of traffic systems and the constantly changing nature of many impacting factors. In this paper, we focus on the spatio-temporal factors, and propose a graph multi-attention network (GMAN) to predict traffic conditions for time steps ahead at different locations on a road network graph. GMAN adapts an encoder-decoder architecture, where both the encoder and the decoder consist of multiple spatio-temporal attention blocks to model the impact of the spatio-temporal factors on traffic conditions. The encoder encodes the input traffic features and the decoder predicts the output sequence. Between the encoder and the decoder, a transform attention layer is applied to convert the encoded traffic features to generate the sequence representations of future time steps as the input of the decoder. The transform attention mechanism models the direct relationships between historical and future time steps that helps to alleviate the error propagation problem among prediction time steps. Experimental results on two real-world traffic prediction tasks (i.e., traffic volume prediction and traffic speed prediction) demonstrate the superiority of GMAN. In particular, in the 1 hour ahead prediction, GMAN outperforms state-of-the-art methods by up to 4% improvement in MAE measure. The source code is available at https://github.com/zhengchuanpan/GMAN.

Xun Huang, Hai Wu, Xin Li et al.

LiDAR-based 3D object detection models have traditionally struggled under rainy conditions due to the degraded and noisy scanning signals. Previous research has attempted to address this by simulating the noise from rain to improve the robustness of detection models. However, significant disparities exist between simulated and actual rain-impacted data points. In this work, we propose a novel rain simulation method, termed DRET, that unifies Dynamics and Rainy Environment Theory to provide a cost-effective means of expanding the available realistic rain data for 3D detection training. Furthermore, we present a Sunny-to-Rainy Knowledge Distillation (SRKD) approach to enhance 3D detection under rainy conditions. Extensive experiments on the WaymoOpenDataset large-scale dataset show that, when combined with the state-of-the-art DSVT model and other classical 3D detectors, our proposed framework demonstrates significant detection accuracy improvements, without losing efficiency. Remarkably, our framework also improves detection capabilities under sunny conditions, therefore offering a robust solution for 3D detection regardless of whether the weather is rainy or sunny

Haoze Wu, Cheng Wang, Wenshuo Zhao et al.

Recent advances in applying reinforcement learning (RL) to large language models (LLMs) have led to substantial progress. In particular, a series of remarkable yet often counterintuitive phenomena have been reported in LLMs, exhibiting patterns not typically observed in traditional RL settings. For example, notable claims include that a single training example can match the performance achieved with an entire dataset, that the reward signal does not need to be very accurate, and that training solely with negative samples can match or even surpass sophisticated reward-based methods. However, the precise conditions under which these observations hold - and, critically, when they fail - remain unclear. In this work, we identify a key factor that differentiates RL observations: whether the pretrained model already exhibits strong Model-Task Alignment, as measured by pass@k accuracy on the evaluated task. Through a systematic and comprehensive examination of a series of counterintuitive claims, supported by rigorous experimental validation across different model architectures and task domains, our findings show that while standard RL training remains consistently robust across settings, many of these counterintuitive results arise only when the model and task already exhibit strong model-task alignment. In contrast, these techniques fail to drive substantial learning in more challenging regimes, where standard RL methods remain effective.

Pufan Zou, Shijia Zhao, Weijie Huang et al.

Recently, Visual Foundation Models (VFMs) have shown a remarkable generalization performance in 3D perception tasks. However, their effectiveness in large-scale outdoor datasets remains constrained by the scarcity of accurate supervision signals, the extensive noise caused by variable outdoor conditions, and the abundance of unknown objects. In this work, we propose a novel label-free learning method, Adaptive Label Correction (AdaCo), for 3D semantic segmentation. AdaCo first introduces the Cross-modal Label Generation Module (CLGM), providing cross-modal supervision with the formidable interpretive capabilities of the VFMs. Subsequently, AdaCo incorporates the Adaptive Noise Corrector (ANC), updating and adjusting the noisy samples within this supervision iteratively during training. Moreover, we develop an Adaptive Robust Loss (ARL) function to modulate each sample's sensitivity to noisy supervision, preventing potential underfitting issues associated with robust loss. Our proposed AdaCo can effectively mitigate the performance limitations of label-free learning networks in 3D semantic segmentation tasks. Extensive experiments on two outdoor benchmark datasets highlight the superior performance of our method.

Dingrui Wang, Zhexiao Sun, Zhouheng Li et al.

A major challenge in deploying world models is the trade-off between size and performance. Large world models can capture rich physical dynamics but require massive computing resources, making them impractical for edge devices. Small world models are easier to deploy but often struggle to learn accurate physics, leading to poor predictions. We propose the Physics-Informed BEV World Model (PIWM), a compact model designed to efficiently capture physical interactions in bird's-eye-view (BEV) representations. PIWM uses Soft Mask during training to improve dynamic object modeling and future prediction. We also introduce a simple yet effective technique, Warm Start, for inference to enhance prediction quality with a zero-shot model. Experiments show that at the same parameter scale (400M), PIWM surpasses the baseline by 60.6% in weighted overall score. Moreover, even when compared with the largest baseline model (400M), the smallest PIWM (130M Soft Mask) achieves a 7.4% higher weighted overall score with a 28% faster inference speed.

Xinzhu Ma, Cheng Wang, Chen Tang et al.

Recovering CAD models from point clouds requires reconstructing their topology and sketch-based extrusion primitives. A dominant paradigm for representing sketches involves implicit neural representations such as Signed Distance Fields (SDFs). However, this indirect approach inherently struggles with precision, leading to unintended curved edges and models that are difficult to edit. In this paper, we propose Point2Primitive, a framework that learns to directly predict the explicit, parametric primitives of CAD models. Our method treats sketch reconstruction as a set prediction problem, employing a improved transformer-based decoder with explicit position queries to directly detect and predict the fundamental sketch curves (i.e., type and parameter) from the point cloud. Instead of approximating a continuous field, we formulate curve parameters as explicit position queries, which are optimized autoregressively to achieve high accuracy. The overall topology is rebuilt via extrusion segmentation. Extensive experiments demonstrate that this direct prediction paradigm significantly outperforms implicit methods in both primitive accuracy and overall geometric fidelity.

Shangbin Wu, Xu Yan, Xiaoliang Fan et al.

Learning the embeddings for urban regions from human mobility data can reveal the functionality of regions, and then enables the correlated but distinct tasks such as crime prediction. Human mobility data contains rich but abundant information, which yields to the comprehensive region embeddings for cross domain tasks. In this paper, we propose multi-graph fusion networks (MGFN) to enable the cross domain prediction tasks. First, we integrate the graphs with spatio-temporal similarity as mobility patterns through a mobility graph fusion module. Then, in the mobility pattern joint learning module, we design the multi-level cross-attention mechanism to learn the comprehensive embeddings from multiple mobility patterns based on intra-pattern and inter-pattern messages. Finally, we conduct extensive experiments on real-world urban datasets. Experimental results demonstrate that the proposed MGFN outperforms the state-of-the-art methods by up to 12.35% improvement.

Chen Zhao, Shi Shi, Zhuo He et al.

Background. Functional assessment of right ventricle (RV) using gated myocardial perfusion single-photon emission computed tomography (MPS) heavily relies on the precise extraction of right ventricular contours. In this paper, we present a new deep-learning-based model integrating both the spatial and temporal features in gated MPS images to perform the segmentation of the RV epicardium and endocardium. Methods. By integrating the spatial features from each cardiac frame of the gated MPS and the temporal features from the sequential cardiac frames of the gated MPS, we developed a Spatial-Temporal V-Net (ST-VNet) for automatic extraction of RV endocardial and epicardial contours. In the ST-VNet, a V-Net is employed to hierarchically extract spatial features, and convolutional long-term short-term memory (ConvLSTM) units are added to the skip-connection pathway to extract the temporal features. The input of the ST-VNet is ECG-gated sequential frames of the MPS images and the output is the probability map of the epicardial or endocardial masks. A Dice similarity coefficient (DSC) loss which penalizes the discrepancy between the model prediction and the ground truth was adopted to optimize the segmentation model. Results. Our segmentation model was trained and validated on a retrospective dataset with 45 subjects, and the cardiac cycle of each subject was divided into 8 gates. The proposed ST-VNet achieved a DSC of 0.8914 and 0.8157 for the RV epicardium and endocardium segmentation, respectively. The mean absolute error, the mean squared error, and the Pearson correlation coefficient of the RV ejection fraction (RVEF) between the ground truth and the model prediction were 0.0609, 0.0830, and 0.6985. Conclusion. Our proposed ST-VNet is an effective model for RV segmentation. It has great promise for clinical use in RV functional assessment.

Xu Yan, Xiaoliang Fan, Peizhen Yang et al.

Representation learning on temporal interaction graphs (TIG) is to model complex networks with the dynamic evolution of interactions arising in a broad spectrum of problems. Existing dynamic embedding methods on TIG discretely update node embeddings merely when an interaction occurs. They fail to capture the continuous dynamic evolution of embedding trajectories of nodes. In this paper, we propose a two-module framework named ConTIG, a continuous representation method that captures the continuous dynamic evolution of node embedding trajectories. With two essential modules, our model exploit three-fold factors in dynamic networks which include latest interaction, neighbor features and inherent characteristics. In the first update module, we employ a continuous inference block to learn the nodes' state trajectories by learning from time-adjacent interaction patterns between node pairs using ordinary differential equations. In the second transform module, we introduce a self-attention mechanism to predict future node embeddings by aggregating historical temporal interaction information. Experiments results demonstrate the superiority of ConTIG on temporal link prediction, temporal node recommendation and dynamic node classification tasks compared with a range of state-of-the-art baselines, especially for long-interval interactions prediction.

Yanfei Su, Weiquan Liu, Zhimin Yuan et al.

Semantic segmentation of building facade is significant in various applications, such as urban building reconstruction and damage assessment. As there is a lack of 3D point clouds datasets related to the fine-grained building facade, we construct the first large-scale building facade point clouds benchmark dataset for semantic segmentation. The existing methods of semantic segmentation cannot fully mine the local neighborhood information of point clouds. Addressing this problem, we propose a learnable attention module that learns Dual Local Attention features, called DLA in this paper. The proposed DLA module consists of two blocks, including the self-attention block and attentive pooling block, which both embed an enhanced position encoding block. The DLA module could be easily embedded into various network architectures for point cloud segmentation, naturally resulting in a new 3D semantic segmentation network with an encoder-decoder architecture, called DLA-Net in this work. Extensive experimental results on our constructed building facade dataset demonstrate that the proposed DLA-Net achieves better performance than the state-of-the-art methods for semantic segmentation.

Xuelun Shen, Qian Hu, Xin Li et al.

This paper presents a matching network to establish point correspondence between images. We propose a Multi-Arm Network (MAN) to learn region overlap and depth, which can greatly improve the keypoint matching robustness while bringing little computational cost during the inference stage. Another design that makes this framework different from many existing learning based pipelines that require re-training when a different keypoint detector is adopted, our network can directly work with different keypoint detectors without such a time-consuming re-training process. Comprehensive experiments conducted on outdoor and indoor datasets demonstrated that our proposed MAN outperforms state-of-the-art methods.

Renfei Tu, Yang Liu, Yongzeng Xue et al.

Being one of the most popular generative framework, variational autoencoders(VAE) are known to suffer from a phenomenon termed posterior collapse, i.e. the latent variational distributions collapse to the prior, especially when a strong decoder network is used. In this work, we analyze the latent representation of collapsed VAEs, and proposed a novel model, neighbor embedding VAE(NE-VAE), which explicitly constraints the encoder to encode inputs close in the input space to be close in the latent space. We observed that for VAE variants that report similar ELBO, KL divergence or even mutual information scores may still behave quite differently in the latent organization. In our experiments, NE-VAE can produce qualitatively different latent representations with majority of the latent dimensions remained active, which may benefit downstream latent space optimization tasks. NE-VAE can prevent posterior collapse to a much greater extent than it's predecessors, and can be easily plugged into any autoencoder framework, without introducing addition model components and complex training routines.

Xian Sun, Peijin Wang, Zhiyuan Yan et al.

With the rapid development of deep learning, many deep learning-based approaches have made great achievements in object detection task. It is generally known that deep learning is a data-driven method. Data directly impact the performance of object detectors to some extent. Although existing datasets have included common objects in remote sensing images, they still have some limitations in terms of scale, categories, and images. Therefore, there is a strong requirement for establishing a large-scale benchmark on object detection in high-resolution remote sensing images. In this paper, we propose a novel benchmark dataset with more than 1 million instances and more than 15,000 images for Fine-grAined object recognItion in high-Resolution remote sensing imagery which is named as FAIR1M. All objects in the FAIR1M dataset are annotated with respect to 5 categories and 37 sub-categories by oriented bounding boxes. Compared with existing detection datasets dedicated to object detection, the FAIR1M dataset has 4 particular characteristics: (1) it is much larger than other existing object detection datasets both in terms of the quantity of instances and the quantity of images, (2) it provides more rich fine-grained category information for objects in remote sensing images, (3) it contains geographic information such as latitude, longitude and resolution, (4) it provides better image quality owing to a careful data cleaning procedure. To establish a baseline for fine-grained object recognition, we propose a novel evaluation method and benchmark fine-grained object detection tasks and a visual classification task using several State-Of-The-Art (SOTA) deep learning-based models on our FAIR1M dataset. Experimental results strongly indicate that the FAIR1M dataset is closer to practical application and it is considerably more challenging than existing datasets.

Yan Xia, Yusheng Xu, Cheng Wang et al.

As a dynamic and essential component in the road environment of urban scenarios, vehicles are the most popular investigation targets. To monitor their behavior and extract their geometric characteristics, an accurate and instant measurement of vehicles plays a vital role in traffic and transportation fields. Point clouds acquired from the mobile laser scanning (MLS) system deliver 3D information of road scenes with unprecedented detail. They have proven to be an adequate data source in the fields of intelligent transportation and autonomous driving, especially for extracting vehicles. However, acquired 3D point clouds of vehicles from MLS systems are inevitably incomplete due to object occlusion or self-occlusion. To tackle this problem, we proposed a neural network to synthesize complete, dense, and uniform point clouds for vehicles from MLS data, named Vehicle Points Completion-Net (VPC-Net). In this network, we introduce a new encoder module to extract global features from the input instance, consisting of a spatial transformer network and point feature enhancement layer. Moreover, a new refiner module is also presented to preserve the vehicle details from inputs and refine the complete outputs with fine-grained information. Given sparse and partial point clouds as inputs, the network can generate complete and realistic vehicle structures and keep the fine-grained details from the partial inputs. We evaluated the proposed VPC-Net in different experiments using synthetic and real-scan datasets and applied the results to 3D vehicle monitoring tasks. Quantitative and qualitative experiments demonstrate the promising performance of the proposed VPC-Net and show state-of-the-art results.

Zhuang Yang, Zengping Chen, Cheng Wang

StochAstic Recursive grAdient algoritHm (SARAH), originally proposed for convex optimization and also proven to be effective for general nonconvex optimization, has received great attention due to its simple recursive framework for updating stochastic gradient estimates. The performance of SARAH significantly depends on the choice of step size sequence. However, SARAH and its variants often employ a best-tuned step size by mentor, which is time consuming in practice. Motivated by this gap, we proposed a variant of the Barzilai-Borwein (BB) method, referred to as the Random Barzilai-Borwein (RBB) method, to calculate step size for SARAH in the mini-batch setting, thereby leading to a new SARAH method: MB-SARAH-RBB. We prove that MB-SARAH-RBB converges linearly in expectation for strongly convex objective functions. We analyze the complexity of MB-SARAH-RBB and show that it is better than the original method. Numerical experiments on standard data sets indicate that MB-SARAH-RBB outperforms or matches state-of-the-art algorithms.

Jonathan Li, Rongren Wu, Yiping Chen et al.

Middle-echo, which covers one or a few corresponding points, is a specific type of 3D point cloud acquired by a multi-echo laser scanner. In this paper, we propose a novel approach for automatic segmentation of trees that leverages middle-echo information from LiDAR point clouds. First, using a convolution classification method, the proposed type of point clouds reflected by the middle echoes are identified from all point clouds. The middle-echo point clouds are distinguished from the first and last echoes. Hence, the crown positions of the trees are quickly detected from the huge number of point clouds. Second, to accurately extract trees from all point clouds, we propose a 3D deep learning network, PointNLM, to semantically segment tree crowns. PointNLM captures the long-range relationship between the point clouds via a non-local branch and extracts high-level features via max-pooling applied to unordered points. The whole framework is evaluated using the Semantic 3D reduced-test set. The IoU of tree point cloud segmentation reached 0.864. In addition, the semantic segmentation network was tested using the Paris-Lille-3D dataset. The average IoU outperformed several other popular methods. The experimental results indicate that the proposed algorithm provides an excellent solution for vegetation segmentation from LiDAR point clouds.

Xuelun Shen, Cheng Wang, Xin Li et al.

This paper proposes a new end-to-end trainable matching network based on receptive field, RF-Net, to compute sparse correspondence between images. Building end-to-end trainable matching framework is desirable and challenging. The very recent approach, LF-Net, successfully embeds the entire feature extraction pipeline into a jointly trainable pipeline, and produces the state-of-the-art matching results. This paper introduces two modifications to the structure of LF-Net. First, we propose to construct receptive feature maps, which lead to more effective keypoint detection. Second, we introduce a general loss function term, neighbor mask, to facilitate training patch selection. This results in improved stability in descriptor training. We trained RF-Net on the open dataset HPatches, and compared it with other methods on multiple benchmark datasets. Experiments show that RF-Net outperforms existing state-of-the-art methods.

Rahul Anand, Hye-Young Paik, Cheng Wang

Large amount of public data produced by enterprises are in semi-structured PDF form. Tabular data extraction from reports and other published data in PDF format is of interest for various data consolidation purposes such as analysing and aggregating financial reports of a company. Queries into the structured tabular data in PDF format are normally processed in an unstructured manner through means like text-match. This is mainly due to that the binary format of PDF documents is optimized for layout and rendering and do not have great support for automated parsing of data. Moreover, even the same table type in PDF files varies in schema, row or column headers, which makes it difficult for a query plan to cover all relevant tables. This paper proposes a deep learning based method to enable SQL-like query and analysis of financial tables from annual reports in PDF format. This is achieved through table type classification and nearest row search. We demonstrate that using word embedding trained on Google news for header match clearly outperforms the text-match based approach in traditional database. We also introduce a practical system that uses this technology to query and analyse finance tables in PDF documents from various sources.

Shenlong Lou, Yan Luo, Qiancong Fan et al.

It is widely recognized that the deeper networks or networks with more feature maps have better performance. Existing studies mainly focus on extending the network depth and increasing the feature maps of networks. At the same time, horizontal expansion network (e.g. Inception Model) as an alternative way to improve network performance has not been fully investigated. Accordingly, we proposed NeuroTreeNet (NTN), as a new horizontal extension network through the combination of random forest and Inception Model. Based on the tree structure, in which each branch represents a network and the root node features are shared to child nodes, network parameters are effectively reduced. By combining all features of leaf nodes, even less feature maps achieved better performance. In addition, the relationship between tree structure and the performance of NTN was investigated in depth. Comparing to other networks (e.g. VDSR\_5) with equal magnitude parameters, our model showed preferable performance in super resolution reconstruction task.

Zongliang Zhang, Hongbin Zeng, Jonathan Li et al.

This paper deals with the geometric multi-model fitting from noisy, unstructured point set data (e.g., laser scanned point clouds). We formulate multi-model fitting problem as a sequential decision making process. We then use a deep reinforcement learning algorithm to learn the optimal decisions towards the best fitting result. In this paper, we have compared our method against the state-of-the-art on simulated data. The results demonstrated that our approach significantly reduced the number of fitting iterations.

Shanxin Zhang, Cheng Wang, Zhuang Yang et al.

The timely provision of traffic sign information to drivers is essential for the drivers to respond, to ensure safe driving, and to avoid traffic accidents in a timely manner. We proposed a timely visual recognizability quantitative evaluation method for traffic signs in large-scale transportation environments. To achieve this goal, we first address the concept of a visibility field to reflect the visible distribution of three-dimensional (3D) space and construct a traffic sign Visibility Evaluation Model (VEM) to measure the traffic sign visibility for a given viewpoint. Then, based on the VEM, we proposed the concept of the Visual Recognizability Field (VRF) to reflect the visual recognizability distribution in 3D space and established a Visual Recognizability Evaluation Model (VREM) to measure a traffic sign visual recognizability for a given viewpoint. Next, we proposed a Traffic Sign Timely Visual Recognizability Evaluation Model (TSTVREM) by combining VREM, the actual maximum continuous visual recognizable distance, and traffic big data to measure a traffic sign visual recognizability in different lanes. Finally, we presented an automatic algorithm to implement the TSTVREM model through traffic sign and road marking detection and classification, traffic sign environment point cloud segmentation, viewpoints calculation, and TSTVREM model realization. The performance of our method for traffic sign timely visual recognizability evaluation is tested on three road point clouds acquired by a mobile laser scanning system (RIEGL VMX-450) according to Road Traffic Signs and Markings (GB 5768-1999 in China), showing that our method is feasible and efficient.

Zongliang Zhang, Jonathan Li, Yulan Guo et al.

Geometric model fitting is a fundamental task in computer graphics and computer vision. However, most geometric model fitting methods are unable to fit an arbitrary geometric model (e.g. a surface with holes) to incomplete data, due to that the similarity metrics used in these methods are unable to measure the rigid partial similarity between arbitrary models. This paper hence proposes a novel rigid geometric similarity metric, which is able to measure both the full similarity and the partial similarity between arbitrary geometric models. The proposed metric enables us to perform partial procedural geometric model fitting (PPGMF). The task of PPGMF is to search a procedural geometric model space for the model rigidly similar to a query of non-complete point set. Models in the procedural model space are generated according to a set of parametric modeling rules. A typical query is a point cloud. PPGMF is very useful as it can be used to fit arbitrary geometric models to non-complete (incomplete, over-complete or hybrid-complete) point cloud data. For example, most laser scanning data is non-complete due to occlusion. Our PPGMF method uses Markov chain Monte Carlo technique to optimize the proposed similarity metric over the model space. To accelerate the optimization process, the method also employs a novel coarse-to-fine model dividing strategy to reject dissimilar models in advance. Our method has been demonstrated on a variety of geometric models and non-complete data. Experimental results show that the PPGMF method based on the proposed metric is able to fit non-complete data, while the method based on other metrics is unable. It is also shown that our method can be accelerated by several times via early rejection.

Cheng Wang, Haojin Yang, Christian Bartz et al.

This work presents an end-to-end trainable deep bidirectional LSTM (Long-Short Term Memory) model for image captioning. Our model builds on a deep convolutional neural network (CNN) and two separate LSTM networks. It is capable of learning long term visual-language interactions by making use of history and future context information at high level semantic space. Two novel deep bidirectional variant models, in which we increase the depth of nonlinearity transition in different way, are proposed to learn hierarchical visual-language embeddings. Data augmentation techniques such as multi-crop, multi-scale and vertical mirror are proposed to prevent overfitting in training deep models. We visualize the evolution of bidirectional LSTM internal states over time and qualitatively analyze how our models "translate" image to sentence. Our proposed models are evaluated on caption generation and image-sentence retrieval tasks with three benchmark datasets: Flickr8K, Flickr30K and MSCOCO datasets. We demonstrate that bidirectional LSTM models achieve highly competitive performance to the state-of-the-art results on caption generation even without integrating additional mechanism (e.g. object detection, attention model etc.) and significantly outperform recent methods on retrieval task.