3D Classification Research Articles

Visual assessment of cerebrospinal fluid flow dynamics using 3D T2-weighted SPACE sequence-based classification system.

Flow-related signal void artifacts can be visualized on the T2-weighted (T2W) three-dimensional sampling perfection with application-optimized contrast (3D-SPACE) sequence. Flow void artifacts in the cerebral aqueduct and the fourth ventricle can provide information about cerebrospinal fluid (CSF) flow dynamics. In this study, we aimed to test the performance of the T2W 3D-SPACE sequence in assessing the CSF flow in the aqueduct and/or fourth ventricle. A total of 137 patients (age range = 3-89 years) who underwent CSF flow study were included. The amount of signal loss on T2W 3D-SPACE due to flow in the aqueduct and fourth ventricle was assessed and graded using a 4-point scale of 0 (absence of flow void) to 3 (signal void filling the aqueduct and entire fourth ventricle). A correlation was then sought between the quantitative values obtained by phase-contrast magnetic resonance imaging (PC-MRI) and the amount of signal void in the 3D-SPACE sequence. At the aqueduct level, there was a statistically significant difference in the forward flow velocity and the flow volume among different grades (all P < 0.001). In the grade 3 group, CSF peak systolic flow velocity and mean flow volume were found to be significantly higher than in the other grades (P < 0.001). The mean aqueduct area in the grade 0 group was found to be significantly different from that in the other classes (P < 0.001). The amount of signal loss in the fourth ventricle observed on T2W 3D-SPACE is correlated with the peak systolic velocity and flow volume measured quantitatively in PC-MRI.

3D dataset generation using virtual reality for forest biodiversity

ABSTRACT While forest biodiversity faces a concerning decline, modern technology presents promising avenues for mitigation. However, a critical gap persists in reconciling ecological knowledge with the technical expertise required to use state-of-the-art technologies in 3D data classification. Currently, one main issue is the scarcity of 3D datasets for biodiversity, particularly within the context of machine learning applications. Unlike the straightforward classification of human-made structures, forest environments are uniquely intricate and nuanced due to its inherently complex nature. This study addresses this challenge by introducing a fully automated pipeline for tree stem 3D point cloud segmentation, focussing on a biodiversity indicator: tree-related microhabitats (TreMs). Furthermore, our research advances the field by demonstrating that machine learning models trained with labels generated by our proposed virtual reality (VR) method, Labelling Flora, yield predictions statistically similar to the traditional desktop-based labelling methods. This implies that existing 3D datasets could be augmented using the more rapid approach of VR labelling. Additionally, the findings of this paper demonstrate the potential integration of VR and immersive technology into the 3D labelling workflow, facilitating a quicker and more intuitive labelling process. This could empower users, who are non-familiar with 3D modelling, to contribute their expertise to the segmentation process.

Multiparametric Characterization of Focal Cortical Dysplasia Using 3D MR Fingerprinting.

To develop a multiparametric machine-learning (ML) framework using high-resolution 3 dimensional (3D) magnetic resonance (MR) fingerprinting (MRF) data for quantitative characterization of focal cortical dysplasia (FCD). We included 119 subjects, 33 patients with focal epilepsy and histopathologically confirmed FCD, 60 age- and gender-matched healthy controls (HCs), and 26 disease controls (DCs). Subjects underwent whole-brain 3 Tesla MRF acquisition, the reconstruction of which generated T1 and T2 relaxometry maps. A 3D region of interest was manually created for each lesion, and z-score normalization using HC data was performed. We conducted 2D classification with ensemble models using MRF T1 and T2 mean and standard deviation from gray matter and white matter for FCD versus controls. Subtype classification additionally incorporated entropy and uniformity of MRF metrics, as well as morphometric features from the morphometric analysis program (MAP). We translated 2D results to individual probabilities using the percentage of slices above an adaptive threshold. These probabilities and clinical variables were input into a support vector machine for individual-level classification. Fivefold cross-validation was performed and performance metrics were reported using receiver-operating-characteristic-curve analyses. FCD versus HC classification yielded mean sensitivity, specificity, and accuracy of 0.945, 0.980, and 0.962, respectively; FCD versus DC classification achieved 0.918, 0.965, and 0.939. In comparison, visual review of the clinical magnetic resonance imaging (MRI) detected 48% (16/33) of the lesions by official radiology report. In the subgroup where both clinical MRI and MAP were negative, the MRF-ML models correctly distinguished FCD patients from HCs and DCs in 98.3% of cross-validation trials. Type II versus non-type-II classification exhibited mean sensitivity, specificity, and accuracy of 0.835, 0.823, and 0.83, respectively; type IIa versus IIb classification showed 0.85, 0.9, and 0.87. In comparison, the transmantle sign was present in 58% (7/12) of the IIb cases. The MRF-ML framework presented in this study demonstrated strong efficacy in noninvasively classifying FCD from normal cortex and distinguishing FCD subtypes. ANN NEUROL 2024;96:944-957.

Mandibular second molar impaction: introducing a novel and validated 3D classification system

BackgroundMandibular second molar (M2M) impaction is a clinically significant manifestation of eruption disturbance in dental development. The primary aim of this study was to investigate the impact of the three-dimensional (3D) characterization on clinical and therapeutic decisions for M2M impaction. The secondary aim was to introduce a validated 3D classification system incorporating both surgical and orthodontic parameters.MethodsBidimensional (2D) and 3D radiological records of 15 impacted M2M were collected and deidentified. Ten experienced clinicians (5 oral surgeons;5 orthodontists) categorized each case, first based on 2D records and then with 3D scans. The degree of orthodontic and surgical difficulty in treating impacted M2M was evaluated using a novel classification system based on anatomical and radiological features. The primary outcome variable was the assessment of differences in diagnosis and decision-making protocol using 2D or 3D records, where clinical relevance ranged from 0 to 4. The secondary outcome variable was the validation analysis of the proposed 3D classification system to determine the concordance among the clinicians. Descriptive statistics and multivariable inferential analysis based on Akaike information criterion (AIC) were performed (α = 0.05).Results3D examination allowed a better visualization of M2M impaction with higher clinical relevance for diagnosis of M2M root relationship to alveolar nerve and lingual plate, height to alveolar crest, depth, and inclination relative to the first molar and position relative to the third molar (range:2.69–3.43). The proposed 3D classification of M2M impaction changed clinical decisions regarding surgical-orthodontic approach, biomechanics, patient education, and treatment time estimate (range:2.59–3.33). In the validation analysis of the classification, no evidence of inter- or intra-group (surgeon/orthodontist) bias in score attribution occurred: the model with the minimum AIC was the null model (AIC = 718.04).Conclusion3D evaluation of impacted M2Ms could enhance diagnostic accuracy, and a classification system was proposed and validated by a group of experienced surgeons and orthodontists with high concordance.

Enhancing MeshNet for 3D shape classification with focal and regularization losses

With the development of deep learning and computer vision, an increasing amount of research has focused on applying deep learning models to the recognition and classification of three-dimensional shapes. In classification tasks, differences in sample quantity, feature amount, model complexity, and other aspects among different categories of 3D model data cause significant variations in classification difficulty. However, simple cross-entropy loss is generally used as the loss function, but it is insufficient to address these differences. In this paper, we used MeshNet as the base model and introduced focal loss as a metric for the loss function. Additionally, to prevent deep learning models from developing a preference for specific categories, we incorporated regularization loss. The combined use of focal loss and regularization loss in optimizing the MeshNet model’s loss function resulted in a classification accuracy of up to 92.46%, representing a 0.20% improvement over the original model’s highest accuracy of 92.26%. Furthermore, the average accuracy over the final 50 epochs remained stable at a higher level of 92.01%, reflecting a 0.71% improvement compared to the original MeshNet model’s 91.30%. These results indicate that our method performs better in 3D shape classification task.

Evaluating Cryo-TEM Reconstruction Accuracy of Self-Assembled Polymer Nanostructures.

Cryogenic transmission electron microscopy (cryo-TEM) combined with single particle analysis (SPA) is an emerging imaging approach for soft materials. However, the accuracy of SPA-reconstructed nanostructures, particularly those formed by synthetic polymers, remains uncertain due to potential packing heterogeneity of the nanostructures. In this study, the combination of molecular dynamics (MD) simulations and image simulations is utilized to validate the accuracy of cryo-TEM 3D reconstructions of self-assembled polypeptoid fibril nanostructures. Using CryoSPARC software, image simulations, 2D classifications, ab initio reconstructions, and homogenous refinements are performed. By comparing the results with atomic models, the recovery of molecular details is assessed, heterogeneous structures are identified, and the influence of extraction location on the reconstructions is evaluated. These findings confirm the fidelity of single particle analysis in accurately resolving complex structural characteristics and heterogeneous structures, exhibiting its potential as a valuable tool for detailed structural analysis of synthetic polymers and soft materials.

3DFFL: privacy-preserving Federated Few-Shot Learning for 3D point clouds in autonomous vehicles

This paper presents a comprehensive study of 3D point cloud Federated Few-Shot Learning (3DFFL), focusing on addressing challenges such as limited data availability and privacy concerns in point cloud classification for applications such as autonomous vehicles. We introduce a novel approach that integrates Federated Learning with Few-Shot Learning techniques, with a special emphasis on optimizing network architectures for 3D point cloud data. Our method capitalizes on the strengths of PointNet++ for feature extraction and ProtoNet for classification, all within a federated learning framework to ensure data privacy and collaborative learning. Significantly, the approach is augmented with the use of attention and SoftMax layers, enhancing the feature extraction and classification processes. Extensive experiments on the ModelNet40, ShapeNet, and ScanOnjectNN datasets validate our method’s accuracy and adaptability in handling 3D point cloud classification, especially in privacy-sensitive and collaborative scenarios. This study not only demonstrates the potential of integrating attention mechanisms and SoftMax layers in 3DFFL but also lays a robust foundation for future advancements in this evolving field, particularly in technologies dependent on 3D data processing.

Graph Transformer for 3D point clouds classification and semantic segmentation

Recently, graph-based and Transformer-based deep learning have demonstrated excellent performances on various point cloud tasks. Most of the existing graph-based methods rely on static graph, which take a fixed input to establish graph relations. Moreover, many graph-based methods apply maximizing and averaging to aggregate neighboring features, so that only a single neighboring point affects the feature of centroid or different neighboring points own the same influence on the centroid’s feature, which ignoring the correlation and difference between points. Most Transformer-based approaches extract point cloud features based on global attention and lack the feature learning on local neighbors. To solve the above issues of graph-based and Transformer-based models, we propose a new feature extraction block named Graph Transformer and construct a 3D point cloud learning network called GTNet to learn features of point clouds on local and global patterns. Graph Transformer integrates the advantages of graph-based and Transformer-based methods, and consists of Local Transformer that use intra-domain cross-attention and Global Transformer that use global self-attention. Finally, we use GTNet for shape classification, part segmentation and semantic segmentation tasks in this paper. The experimental results show that our model achieves good learning and prediction ability on most tasks. The source code and pre-trained model of GTNet will be released on https://github.com/NWUzhouwei/GTNet.

Granformer: A granular transformer net with linear complexity

Recently, transformer models have demonstrated excellent performance across various intelligent applications owing to their ability to understand global context through self-attention mechanism. However, the extensively investigated multiplicative-based attention mechanism is inadequate for capturing relationship-based feature representations, as the dot product cannot depict the intricate semantic information between objects. Moreover, it has a great computational burden with a complexity of O(n2), due to the global feature representation capability achieved by calculating the relationship between each token within the entire feature sequence. To solve the current problem, this paper proposes a granular transformer framework with linear complexity, wherein diverse granulation functions can be employed to supersede the prevailing multiplicative relationships, and an innovative linearization methodology in the form of matrix factorization is designed to reduce the computational burden. Relying on the intricate semantics information embedded within granular structures, the capacity for feature extraction is significantly more comprehensive. Then, a novel matrix factorization methodology is developed for the linearity of granulation-based attention, accomplished by implementing separate deformable convolution sampling and using an approximate iterative algorithm based on cubic equations to calculate the Moore–Penrose inverse. The mathematical proof that our method is approximate with the complete granulation-based attention matrix is investigated in detail. Finally, the performance of Granformer, an innovative reconfiguration of plug-and-play transformer block, is evaluated on representative intelligent applications, including 3D point cloud classification, emotion recognition and sentiment analysis, and object detection. The experimental results suggest that our methodologies outperform the state-of-the-art models.

SPOT-RASTR-A cryo-EM specimen preparation technique that overcomes problems with preferred orientation and the air/water interface.

In cryogenic electron microscopy (cryo-EM), specimen preparation remains a bottleneck despite recent advancements. Classical plunge freezing methods often result in issues like aggregation and preferred orientations at the air/water interface. Many alternative methods have been proposed, but there remains a lack a universal solution, and multiple techniques are often required for challenging samples. Here, we demonstrate the use of lipid nanotubes with nickel NTA headgroups as a platform for cryo-EM sample preparation. His-tagged specimens of interest are added to the tubules, and they can be frozen by conventional plunge freezing. We show that the nanotubes protect samples from the air/water interface and promote a wider range of orientations. The reconstruction of average subtracted tubular regions (RASTR) method allows for the removal of the nanotubule signal from the cryo-EM images resulting in isolated images of specimens of interest. Testing with β-galactosidase validates the method's ability to capture particles at lower concentrations, overcome preferred orientations, and achieve near-atomic resolution reconstructions. Since the nanotubules can be identified and targeted automatically at low magnification, the method enables fully automated data collection. Furthermore, the particles on the tubes can be automatically identified and centered using 2D classification enabling particle picking without requiring prior information. Altogether, our approach that we call specimen preparation on a tube RASTR holds promise for overcoming air-water interface and preferred orientation challenges and offers the potential for fully automated cryo-EM data collection and structure determination.

Automatic anal sphincter integrity detection from ultrasound images via convolutional neural networks.

The anal sphincter complex comprises the anal sphincter and the U-shaped deep and superficial puborectalis muscle. As an important supporting structure of the posterior pelvic floor, together with its surrounding tissues and muscles, the anal sphincter complex maintains the normal physiological functions of defecation and continence. The plane required for diagnosing anal sphincter injury and the diagnosis of anal sphincter integrity through pelvic floor ultrasound are highly dependent on sonographers' experience. We developed a deep learning (DL) tool for the automatic diagnosis of anal sphincter integrity via pelvic floor ultrasound. A 2D detection network was trained to detect the bounding box of the anal sphincter. The pelvic floor ultrasound image and its corresponding oval mask were input into a 2D classification network to determine the integrity of the anal sphincter. The average precision (AP) and intersection over union (IoU) were used to evaluate the performance of anal sphincter detection. Receiver operating characteristic (ROC) analysis was used to evaluate the performance of the classification model. The Pearson correlation coefficients (r values) of the topmost and bottommost layers detected by the CNN and sonographers were 0.932 and 0.978, respectively. The best DL model yielded the highest area under the curve (AUC) of 0.808 (95% CI: 0.698-0.921) in the test cohort. The results from the CNN agreed well with the diagnostic results of experienced sonographers. We proposed, for the first time, a CNN to obtain the plane required for diagnosing anal sphincter injury on the basis of pelvic floor ultrasound and for preliminarily diagnosing anal sphincter injury.

The Chalmers Cloud Ice Climatology: retrieval implementation and validation

Abstract. Ice clouds are a crucial component of the Earth's weather system, and their representation remains a principal challenge for current weather and climate models. Several past and future satellite missions were explicitly designed to provide observations offering new insights into cloud processes, but these specialized cloud sensors are limited in their spatial and temporal coverage. Geostationary satellites have been observing clouds for several decades and can ideally complement the sparse measurements from specialized cloud sensors. However, the geostationary observations that are continuously and globally available over the full observation record are restricted to a small number of wavelengths, which limits the information they can provide on clouds. The Chalmers Cloud Ice Climatology (CCIC) is a novel cloud-property dataset that aims to provide an improved climate record of ice hydrometeor concentrations by applying state-of-the-art machine-learning techniques to retrieve ice cloud properties from globally gridded, single-channel geostationary observations that are readily available from 1980 onwards. CCIC offers a novel perspective on the record of geostationary IR observations by providing spatially and temporally continuous retrievals of the vertically integrated and vertically resolved concentrations of frozen hydrometeors, typically referred to as ice water path (IWP) and ice water content (IWC). In addition to that, CCIC provides 2D and 3D cloud masks and a 3D cloud classification. A fully convolutional quantile regression neural network constitutes the core of the CCIC retrieval, providing probabilistic estimates of IWP and IWC. The network is trained against CloudSat retrievals using 3.5 years of global collocations. Assessed on a held-out test dataset, the CCIC-provided IWP and IWC estimates achieve correlations exceeding 0.7 and 0.6, respectively, and biases better than −5 % and −2 % demonstrating considerable skill in estimating both IWP and IWC. In addition, CCIC is extensively validated against both in situ and remote sensing measurements from two flight campaign series and a ground-based radar. The results of this independent validation confirm the ability of CCIC to retrieve IWP and IWC. CCIC thus ideally complements temporally and spatially more limited measurements from dedicated cloud sensors by providing spatially and temporally continuous estimates of ice cloud properties. The CCIC network and its associated software are made accessible to the scientific community.

P-269 Three-dimensional assessment of the inner cell mass is more effective at predicting implantation and live birth than Gardner grading

Abstract Study question 3-dimensional quantitative assessment of the inner cell mass (ICM) is more effective at selecting viable embryos than the widely used Gardner grading system? Summary answer 3-dimensional quantitative assessment of the ICM is significantly more predictive of implantation and live birth than the widely used Gardner grading of the ICM What is known already ICM assessment forms a key part of most widely used blastocyst grading systems. Grading under these systems, however, is often qualitative, subjective, and subject to inter/intra-operator variability. To reduce subjectivity in ICM grading, it has been suggested that quantitative criteria are used. Previous studies have focused on the assessment of such parameters (e.g. ICM area) in 2 dimensions. However, such approaches fail to account for the various embryo orientations which could lead to different measurements. Therefore, quantitative 3-dimensional analysis of the ICM is needed to improve the reliability and consistency of grading. Study design, size, duration This was a retrospective analysis of 153 randomly selected tEB embryos from two clinics. All data was captured on Embryoscope between 2018 and 2020. Data for maternal age, blastocyst grade, blastocyst ploidy, implantation and live birth rates were also recorded. Participants/materials, setting, methods Embryos were imaged at 7+ focal planes and in-focus sections of each ICM were annotated using ImageJ. These sections were combined into a 3D model using Blender, and sphericity (ψ) computed. Embryos were categorised as spheroid (ψ ≥ 0.8), ellipsoid (ψ &amp;lt; 0.8), crescent and irregular. Chi-squared tests were performed, comparing classification and clinical outcome. The predictive power of the system was compared to the Gardner scale (spheroid/ellipsoid classifications and A grades considered predictive of a positive outcome respectively). Main results and the role of chance Most ICM were either ellipsoid (48%) and spheroid (37%) rather than crescent (6%) or irregular (8%, p &amp;lt; 0.001). ICM 3D classification was associated with implantation, X2(3, N = 119)= 11.10, p = 0.01 and live birth, X2(3, N = 119)= 10.78, p = 0.01. Blastocysts with ellipsoid (70%, 28/40) and spheroid (58%, 18/31) shaped ICM had the greatest chances of implantation, whilst those with crescent (40%, 6/15) and irregularly shaped ICM (33%, 11/33) had lower chances of implantation (p &amp;lt; 0.011). Similarly, blastocysts with ellipsoid (58%, 23/40) and spheroid (48%, 15/31) shaped ICM had the greatest chances of live birth, whilst those with crescent (33%, 5/15) and irregularly shaped ICM (21%, 7/33) had lower chances of live birth (p &amp;lt; 0.013). 3D assessment of ICM outperformed Gardner ICM grading with regards to prediction of implantation (accuracy 64.7% vs 49.6%, area-under-the-curve 0.642 vs 0.486, precision 64.8% vs 51.9%, sensitivity 73% vs 65.1%, specificity 55.4% vs 32.1%, F1 score 68.7% vs 57.8%, 3D vs Gardner) and live birth (accuracy 62.2% vs 50.4%, area-under-the-curve 0.641 vs 0.531, precision 53.5% vs 44.3%, sensitivity 76% vs 70%, specificity 52% vs 36.2%, F1 score 62.2% vs 54.3%, 3D vs Gardner). Limitations, reasons for caution Manually rendering the 3D images for each ICM was time consuming, making it impractical for clinical application. Further work is now focusing on automating this process using image recognition and augmented reality. Wider implications of the findings This is a novel ICM assessment that provides an improved perspective of the 3-dimensionality of the ICM which is more consistent and effective than traditional methods of embryo assessment, potentially contributing towards improving embryo selection efficacy and time to pregnancy. Trial registration number none

A cross-species transcriptomic analysis reveals a novel 2-dimensional classification system explaining the invasiveness heterogeneity of pancreatic neuroendocrine tumor

Pancreatic neuroendocrine tumors (PanNETs), the second most common type of primary pancreatic tumors, display notable heterogeneity in invasiveness. Current knowledge regarding genomic alterations, including DAXX/ATRX, MEN1 mutations, and copy number variations (CNVs), provides some insights into tumor invasiveness. However, the underlying reasons for the significant variation in invasiveness between insulinoma and other types of PanNETs remain unclear. To construct a comprehensive model for the stratification of prognosis, we employed analysis of both the well-established Rip1-Tag2 (RT2) mouse model of PanNETs and human PanNETs with various functional types. Firstly, by applying single-cell and bulk RNA sequencing in PanNETs from different ages and strains of RT2 mice and human PanNETs, we introduced a 2-dimensional (2D) classification system. Based on the 2-D classification system, human PanNETs were mainly classified as benign insulinomas or non-insulinomas subclusters. Non-insulinomas subtypes mainly included gastrinomas, glucagonomas, VIPomas, and NF-PanNETs, which all exhibited potential invasiveness. In addition, we discovered an enrichment of specific CNV patterns and mutations in corresponding human PanNET subclusters. Then we denoted somatic DAXX/ATRX as the ‘second hit’ and confounding factors for invasiveness. Finally, by combining the 2D system, DAXX/ATRX mutation status, and tumor diameter, a group of indolent PanNETs with minimal recurrence risk was identified. In conclusion, our current work constructed a comprehensive model to elucidate the heterogeneity of invasiveness in PanNETs and improve prognostic stratification.

Fuzzy preference matroids rough sets for approximate guided representation in transformer

Recently, the transformer has exhibited remarkable performance across various applications, primarily owing to its exceptional capability in capturing global information through the attention mechanism. Nevertheless, the dot-product within attention results in inaccurate and confusing guidance in the presence of uncertain features, subsequently affecting robustness. Hence, introducing an approximate guided method to handle uncertain features is essential for achieving enhanced objectivity. In this study, we employed fuzzy preference relations to construct a matroids structure, introducing it into rough sets to form the fuzzy preference matroids rough sets. Detailed mathematical properties and axiomatic proofs are presented. The closed sets within the matroids form a subspace of the lower approximation proving that the proposed satisfy the approximation ability of rough sets. Further, an approximate guided representation method with a lower approximation of the matroids has been developed. It is integrated into a plug-and-play transformer block that can be flexibly deployed across various tasks. Experimental validation has demonstrated the outstanding performance of this method across 3D point cloud classification, named entity recognition, multimodal emotion recognition, and image classification. Specifically, on the Weibo NER and the IEMOCAP datasets, attain state-of-the-art (SOTA) F1-scores. The resource code is validated at https://github.com/WinnieSunning/FPMRST.

Beyond local patches: Preserving global–local interactions by enhancing self-attention via 3D point cloud tokenization

Transformer-based architectures have recently shown impressive performance on various point cloud understanding tasks such as 3D object shape classification and semantic segmentation. Particularly, this can be attributed to their self-attention mechanism, which has the ability to capture long-range dependencies. However, current methods have constrained it to operate in local patches due to its quadratic memory constraints. This hinders their generalization ability and scaling capacity due to the loss of non-locality in early layers. To tackle this issue, we propose a window-based transformer architecture that captures long-range dependencies while aggregating information in the local patches. We do this by interacting each window with a set of global point cloud tokens — a representative subset of the entire scene — and augmenting the local geometry through a 3D Histogram of Oriented Gradients (HOG) descriptor. Through a series of experiments on segmentation and classification tasks, we show that our model exceeds the state-of-the-art on S3DIS semantic segmentation (+1.67% mIoU), ShapeNetPart part segmentation (+1.03% instance mIoU) and performs competitively on ScanObjectNN 3D object classification.11The code and trained models shall be made publicly available.

ModelNet-O: A large-scale synthetic dataset for occlusion-aware point cloud classification

Recently, 3D point cloud classification has made significant progress with the help of many datasets. However, these datasets do not reflect the incomplete nature of real-world point clouds caused by occlusion, which limits the practical application of current methods. To bridge this gap, we propose ModelNet-O, a large-scale synthetic dataset of 123,041 samples that emulates real-world point clouds with self-occlusion caused by scanning from monocular cameras. ModelNet-O is 10 times larger than existing datasets and offers more challenging cases to evaluate the robustness of existing methods. Our observation on ModelNet-O reveals that well-designed sparse structures can preserve structural information of point clouds under occlusion, motivating us to propose a robust point cloud processing method that leverages a critical point sampling (CPS) strategy in a multi-level manner. We term our method PointMLS. Through extensive experiments, we demonstrate that our PointMLS achieves state-of-the-art results on ModelNet-O and competitive results on regular datasets such as ModelNet40 and ScanObjectNN, and we also demonstrate its robustness and effectiveness. Code available: https://github.com/fanglaosi/ModelNet-O_PointMLS.

A rapid evaluation method of the seismic damage to buildings based on UAV images

With the widespread application of UAVs in post-disaster damage evaluation, feature extraction and analysis based on drone images has become the main trend in seismic assessment. In this research, a fast assessment method that integrates object-oriented 2D classification and 3D assessment is proposed in the paper. Firstly, to segment the regional orthoimage, a multi-scale feature segmentation method is developed to segment the orthoimage of the region and the features are selected based on the attributes of ground objects. Meanwhile, a two-stage object-oriented hierarchical classification method is put forward to establish partition rules at different levels. The classification method takes the overall attributes of the object into account which reduces the amount of data correspondingly resulting in much higher efficiency and accuracy of feature extraction than the pixel-based method. In addition, the integrity of building’s top texture after classification is used as the first criterion to assess building seismic damage. Besides, a 3D elevation analysis method based on regional normalized digital surface model (nDSM) is developed by comparing digital surface model with digital elevation model to serve as the second criterion for determining the vertical changes in building. The proposed method is validated by a field test at the Beichuan earthquake site with an accuracy rate of 86.95 %. The experimental results show that the proposed method can meet the accuracy requirement of seismic assessment and has a high application value in the field of regional disaster evaluation.

ESTATE: A Large Dataset of Under-Represented Urban Objects for 3D Point Cloud Classification

Abstract. Cityscapes contain a variety of objects, each with a particular role in urban administration and development. With the rapid growth and implementation of 3D imaging technology, urban areas are increasingly surveyed with high-resolution point clouds. This technical advancement extensively improves our ability to capture and analyse urban environments and their small objects. Deep learning algorithms for point cloud data have shown considerable capacity in 3D object classification but still face problems with generally under-represented objects (such as light poles or chimneys). This paper introduces the ESTATE dataset (https://github.com/3DOM-FBK/ESTATE), which combines available datasets of various sensors, densities, regions, and object types. It includes 13 classes featuring intensity and/or colour attributes. Tests using ESTATE demonstrate that the dataset improves the classification performance of deep learning techniques and could be a game-changer to advance in the 3D classification of urban objects.

3D documentation and classification of incense tree Aquilaria sinensis (Lour.) Spreng. wounds by photogrammetry and its potential conservation applications.

In recent years, illegal felling of and damage to the incense tree Aquilaria sinensis (Lour.) Spreng. have been reported in Hong Kong. Their native populations are under increasingly severe threat. Therefore, the development of a standard and efficient method to classify and document wounds on vulnerable trees is urgently needed for conservation purposes. In this study, photogrammetry was used to document wounds in A. sinensis through 3D modeling. A total of 752 wound records from 484 individual A. sinensis trees from Hong Kong were included to establish a new wound classification system. Our major findings include a novel standardized procedure for photogrammetric documentation and a wound classification system. The results of this study will facilitate A. sinensis conservation, by enhancing wound documentation and information transfer to law enforcement and education.