Explore the vast range of titles available.

Point cloud analysis has a wide range of applications in many areas such as computer vision, robotic manipulation, and autonomous driving. While deep learning has achieved remarkable success on image-based tasks, there are many unique challenges faced by deep neural networks in processing massive, unordered, irregular and noisy 3D points. To stimulate future research, this paper analyzes recent progress in deep learning methods employed for point cloud processing and presents challenges and potential directions to advance this field. It serves as a comprehensive review on two major tasks in 3D point cloud processing—namely, 3D shape classification and semantic segmentation.

This paper reported a study on the 3-dimensional deep-learning-based automatic diagnosis of nasal fractures. (1) Background: The nasal bone is the most protuberant feature of the face; therefore, it is highly vulnerable to facial trauma and its fractures are known as the most common facial fractures worldwide. In addition, its adhesion causes rapid deformation, so a clear diagnosis is needed early after fracture onset. (2) Methods: The collected computed tomography images were reconstructed to isotropic voxel data including the whole region of the nasal bone, which are represented in a fixed cubic volume. The configured 3-dimensional input data were then automatically classified by the deep learning of residual neural networks (3D-ResNet34 and ResNet50) with the spatial context information using a single network, whose performance was evaluated by 5-fold cross-validation. (3) Results: The classification of nasal fractures with simple 3D-ResNet34 and ResNet50 networks achieved areas under the receiver operating characteristic curve of 94.5% and 93.4% for binary classification, respectively, both indicating unprecedented high performance in the task. (4) Conclusions: In this paper, it is presented the possibility of automatic nasal bone fracture diagnosis using a 3-dimensional Resnet-based single classification network and it will improve the diagnostic environment with future research.

Classification systems for Adolescent Idiopathic Scoliosis (AIS) play an important role in guiding both surgical planning and conservative treatments. Traditional 2D classification systems, such as the Lenke, King and Lehnert‐Schroth classifications, have been widely used for the clinical diagnosis and treatment of scoliosis. However, with the growing understanding of the three‐dimensional nature of scoliosis and advancements in 3D reconstruction technologies, 3D classification systems are gaining increasing attention. This paper reviews the current applications, advantages, and limitations of different 2D and 3D classification systems, focusing on their clinical significance in treatment planning. While 3D classification systems offer clear advantages in capturing the complexity of spinal deformities, their clinical implementation faces challenges such as high costs and technical complexity. Additionally, studies show that computer‐assisted technologies, artificial intelligence can significantly improve the accuracy and consistency of classification systems, reducing human errors. The paper also explores the future directions of classification system development, emphasizing the potential of combining 2D and 3D technologies and the impact of these advancements on personalized scoliosis treatment. This study compares traditional 2D scoliosis classification systems (e.g., King, Lenke, and PUMC) with emerging 3D systems. It highlights the limitations of 2D methods and the advantages of integrating 3D models for more precise treatment planning and outcomes. The analysis shows that while 3D classification offers high reliability and comprehensive insights, challenges like complex modeling and clinical integration remain.

In recent years, Artificial Intelligence (AI) methods for 3D point cloud classification has assumed an essential role in the heritage field. The association of semantic information to 3D representations became a valuable instrument for measurement, analysis, education or maintenance, in particular in the Cultural Heritage (CH) sector. Moreover, the recent availability and reliability of head-mounted displays and glasses are allowing extraordinary immersive virtual experiences. This paper presents an end-to-end framework to handle large and complex 3D point clouds, from acquisition to semantic segmentation and final access in a Mixed Reality (MR) environment. Three completely different heritage scenarios are considered: the Temple of Neptune in Paestum, the Milan Cathedral and a large portion of Bologna’s porticoes. Mixed Reality experiences are described and shown based on the Microsoft HoloLens 2 device. A video of the results is available at https://youtu.be/Kd_3s0tIX04.

Improving the effectiveness of spatial shape features classification from 3D lidar data is very relevant because it is largely used as a fundamental step towards higher level scene understanding challenges of autonomous vehicles and terrestrial robots. In this sense, computing neighborhood for points in dense scans becomes a costly process for both training and classification. This paper proposes a new general framework for implementing and comparing different supervised learning classifiers with a simple voxel-based neighborhood computation where points in each non-overlapping voxel in a regular grid are assigned to the same class by considering features within a support region defined by the voxel itself. The contribution provides offline training and online classification procedures as well as five alternative feature vector definitions based on principal component analysis for scatter, tubular and planar shapes. Moreover, the feasibility of this approach is evaluated by implementing a neural network (NN) method previously proposed by the authors as well as three other supervised learning classifiers found in scene processing methods: support vector machines (SVM), Gaussian processes (GP), and Gaussian mixture models (GMM). A comparative performance analysis is presented using real point clouds from both natural and urban environments and two different 3D rangefinders (a tilting Hokuyo UTM-30LX and a Riegl). Classification performance metrics and processing time measurements confirm the benefits of the NN classifier and the feasibility of voxel-based neighborhood.

Deep learning is robust to the perturbation of a point cloud, which is an important data form in the Internet of Things. However, it cannot effectively capture the local information of the point cloud and recognize the fine-grained features of an object. Different levels of features in the deep learning network are integrated to obtain local information, but this strategy increases network complexity. This paper proposes an effective point cloud encoding method that facilitates the deep learning network to utilize the local information. An axis-aligned cube is used to search for a local region that represents the local information. All of the points in the local region are available to construct the feature representation of each point. These feature representations are then input to a deep learning network. Two well-known datasets, ModelNet40 shape classification benchmark and Stanford 3D Indoor Semantics Dataset, are used to test the performance of the proposed method. Compared with other methods with complicated structures, the proposed method with only a simple deep learning network, can achieve a higher accuracy in 3D object classification and semantic segmentation.

The computer vision, graphics, and machine learning research groups have given a significant amount of focus to 3D object recognition (segmentation, detection, and classification). Deep learning approaches have lately emerged as the preferred method for 3D segmentation problems as a result of their outstanding performance in 2D computer vision. As a result, many innovative approaches have been proposed and validated on multiple benchmark datasets. This study offers an in-depth assessment of the latest developments in deep learning-based 3D object recognition. We discuss the most well-known 3D object recognition models, along with evaluations of their distinctive qualities.

3D point clouds feature valuable geometric and, often, radiometric and semantic information to support studies, analyses and understanding of the surveyed scene. Due to their generally large size, the use and interpretation of point clouds could be problematic. User-friendly and quick approaches for querying these valuable datasets and retrieving information could surely support end-users, in particular in the heritage sector. This work presents an ontology-based approach to facilitate the query and use of 3D heritage point clouds by means of sets of rules in order to infer properties and characteristics of the surveyed scene. Our approach is focused on linking together 3D spatial data and expert knowledge, in a way that the ontology can elaborate, represent, enrich and query a given point cloud. Results show how different queries can be set-up and how the procedure can be replicated to various queries and datasets.

BackgroundThe current classifications of adolescent idiopathic scoliosis (AIS) aim to guide surgical decision making. However, variance exists within treatment recommendations and suboptimal outcomes have been observed while following these guidelines based on two-dimensional images. We used previously developed 3D classification for right thoracic AIS patients and aimed to determine the variation in surgical decision making and the risk of suboptimal outcomes in each subtype according to our classification.MethodsSeventy-six right thoracic AIS patients with 2-year follow-up were included retrospectively. Five 3D preoperative subgroups were determined based on a previous classification system. The upper and lower instrumented vertebrae (UIV and LIV) and the radiographic surgical outcomes at 2-year [frontal balance (FB), proximal junctional kyphosis (PJK), and adding on] were compared between the subtypes.ResultsThe fusion length and the rate of radiographic suboptimal outcomes were statistically different between the five groups. LIV at T12 in Type 1 and UIV at T2 in Type 2 were associated with improved FB and lower PJK, respectively. Type 3 had the highest rate of suboptimal FB and developing PJK. Type 4 had the longest fusion, and suboptimal FB was observed in 42% of the patients independent from the LIV level. Type 5 had the lowest rate of unsatisfactory radiographic outcomes at 2 years.ConclusionFollowing the preoperative 3D classification of the AIS patients, we showed that the UIV and LIV selection has a different impact on the surgical outcomes in each of the five subtypes. The proposed 3D classification has the potential for risk stratification following a posterior spinal surgery in right thoracic AIS.Graphic abstractThese slides can be retrieved under Electronic Supplementary Material.

Objective Existing 3D classification systems for scoliosis primarily guide surgical treatment, with limited application in conservative management. This study aims to establish a preliminary 3D classification system for moderate adolescent idiopathic scoliosis patients in China, providing a theoretical foundation for the standardization and automation of conservative treatment plans. Methods Data from 404 adolescent idiopathic scoliosis patients who did not undergo surgery were retrospectively collected from 2022 to 2025. EOS imaging technology was used to perform 3D reconstruction for each patient. The parameters included the 3D centroid coordinates of the vertebrae and vertebral angular displacement. A total of 102 features were extracted per model, and dimensionality reduction yielded 30 final features by the Stacked Autoencoder method. Fuzzy C‐means clustering with two classification approaches is used: direct clustering and iterative clustering. Iterative clustering was performed based on coronal plane parameters for initial classification, followed by further clustering. Direct classification involved immediate clustering without further subdivision. Results Clustering identified 8 distinct 2D curve types, which were further subdivided into 13 3D subtypes. A comparison of the 13 clusters from direct classification with those obtained from iterative clustering was made using Euclidean and Mahalanobis distances between cluster centers and clinical data. The difference in similarity was higher for direct classification, indicating greater variability. Conclusion EOS imaging technology combined with Fuzzy C‐Means iterative clustering enables a preliminary 3D classification of AIS by capturing more detailed and individualized morphological features. Compared to direct clustering, the iterative method not only improves geometric interpretability but also enhances classification accuracy by better identifying subtle variations in spinal curvature. It further improves specificity, particularly in distinguishing sagittal and axial plane deformities, which are often overlooked in 2D systems. This enhanced resolution provides a stronger basis for developing personalized conservative treatment plans, such as brace design and rehabilitation strategy. Although the proposed method shows promise, further clinical validation is needed to confirm its effectiveness in guiding conservative treatment decisions. This study introduces a novel stepwise C‐means clustering approach for 3D classification of AIS, focusing on conservative treatment. The iterative process enhances subtype identification by integrating coronal and 3D data, offering better clinical understanding, high scalability, and automated parameter acquisition to optimize scoliosis management strategies.