Accurate 3D Research Articles

CardiacField: computational echocardiography for automated heart function estimation using two-dimensional echocardiography probes

Abstract Aims Accurate heart function estimation is vital for detecting and monitoring cardiovascular diseases. While two-dimensional echocardiography (2DE) is widely accessible and used, it requires specialized training, is prone to inter-observer variability, and lacks comprehensive three-dimensional (3D) information. We introduce CardiacField, a computational echocardiography system using a 2DE probe for precise, automated left ventricular (LV) and right ventricular (RV) ejection fraction (EF) estimations, which is especially easy to use for non-cardiovascular healthcare practitioners. We assess the system’s usability among novice users and evaluate its performance against expert interpretations and advanced deep learning (DL) tools. Methods and results We developed an implicit neural representation network to reconstruct a 3D cardiac volume from sequential multi-view 2DE images, followed by automatic segmentation of LV and RV areas to calculate volume sizes and EF values. Our study involved 127 patients to assess EF estimation accuracy against expert readings and two-dimensional (2D) video-based DL models. A subset of 56 patients was utilized to evaluate image quality and 3D accuracy and another 50 to test usability by novice users and across various ultrasound machines. CardiacField generated a 3D heart from 2D echocardiograms with <2 min processing time. The LVEF predicted by our method had a mean absolute error (MAE) of 2.48%, while the RVEF had an MAE of 2.65%. Conclusion Employing a straightforward apical ring scan with a cost-effective 2DE probe, our method achieves a level of EF accuracy for assessing LV and RV function that is comparable to that of three-dimensional echocardiography probes.

A computer-vision based framework for virtual 3D garment reconstruction

AbstractExisting 3D garment reconstruction methods are difficult to implement for online fashion design and e-commerce or special applications. This paper proposes a novel computer-vision framework for 3D garment digital reconstruction, which aims to reconstruct high-quality and realistic virtual 3D garments with fabric mechanic properties for 3D virtual try-on. The new segmentation scheme is proposed to separate the 3D garment point clouds from background points, which is suitable for 3D human shapes and is adaptive for different 3D garment models in different scenes. The new Statistical Outlier Removal algorithm and the learning-based method PointCleanNet are combined to remove noise and outliers in 3D garment point clouds, which provides high-fidelity and high-quality 3D garment point clouds. The 3D garment meshes are then reconstructed from their corresponding point clouds with a modified rolling ball algorithm. Finally, the meshes are improved and converted into physics-based virtual try-on 3D garments with fabric mechanic properties added, which enables the assessment of different body shapes with varied sizes for the same reconstructed 3D garment. Comparison experiments demonstrate that our framework achieves high-quality and realistic 3D garment reconstruction and accurate 3D virtual try-on from 2D garment images. We also demonstrate the proposed framework on a large range of various garments to show this approach has a great potential for garment future technology, such as online garment shopping, garment design and manufacturing.

A Novel Technique for Fluorescence Lifetime Tomography.

Fluorescence lifetime has emerged as a unique imaging modality for quantitatively assessing in vivo the molecular environment of diseased tissues. Although fluorescence lifetime microscopy (in 2D) is a mature field, 3D imaging in deep tissues remains elusive and challenging owing to scattering. Herein, we report on a deep neural network (coined AUTO-FLI) that performs both 3D intensity and quantitative lifetime reconstructions in deep tissues. The proposed Deep Learning (DL)-based approach involves an in silico scheme to generate fluorescence lifetime data accurately. The developed DL model is validated both in silico and on experimental phantoms. Overall, AUTO-FLI provides accurate 3D quantitative estimates of both intensity and lifetime distributions in highly scattering media, demonstrating its unique potential for fluorescence lifetime-based molecular imaging at the mesoscopic and macroscopic scale.

Three-dimensional reconstruction of porous media by fusing multi-grid image features based on extended feature pyramid network

Microstructural analysis of porous media has significant research value in the study of their macroscopic properties. An accurate reconstruction of the digital microstructure model is a prerequisite for understanding their physical properties and transport properties. Deep learning can accurately extract features from training images for a fast and precise reconstruction of porous media. The reconstruction methods based on deep learning have yielded important achievements in the reconstruction of porous media. Aiming at the complex morphology and multi-scale image information of porous media, this paper proposes a generic end-to-end deep learning-based encoder–decoder generative method for the reconstruction of porous media from a single 2D image. This paper designs an extended feature pyramid network encoder to extract image features of different resolutions and different scales, and a generator network incorporating multi-scale image information. Through the training process of the network, the final three-dimensional images that meets certain requirements are generated by means of confrontation. By introducing the loss function of the pattern density function and the loss function of Wasserstein distance with gradient penalty, we constrain the reconstructed 3D structures. Finally, the reconstructed 3D structures and the target 3D structures are visualized to compare their similar morphologies. The morphological parameter indicators of the reconstructed 3D structures were line with those of the target 3D structure. The microstructure parameters of the 3D structures were also compared and analyzed. For complex porous media images, our method can achieve diverse, accurate, and stable 3D reconstruction.

Progressively global–local fusion with explicit guidance for accurate and robust 3d hand pose reconstruction

Parametric and non-parametric methods are two commonly used strategies in current 3D hand pose reconstruction. Parametric methods predict low-dimensional parameters to fit a predefined hand model to the input image. Benefiting from the prior knowledge of hand models, parametric methods guarantee plausible hand poses, whereas the pose estimation accuracy is limited due to nonlinear regression and spatial information loss. Differently, non-parametric methods directly estimate the coordinates of keypoints or mesh vertices from the input image. The reconstructed 3D hand poses show high precision but may be less robust. In this paper, we integrate the advantages of two methods for accurate and robust hand pose reconstruction. Specifically, we disentangle the hand pose reconstruction into global modeling and local refinement, which is performed in a coarse-to-fine manner. Firstly, we utilize global features from the encoder to generate the initial estimation by a parametric method, which aims to provide the prior knowledge of the human hand for subsequent processes. Then, we gradually fuse multi-scale contextual features for local refinement by explicitly integrating global prior information and local visual features. In particular, we introduce a consecutive pixel-aligned feature retrieval module to extract fine-grained information from visual features, thereby achieving pixel-level alignment. Furthermore, we demonstrate that our method can be extended to weakly-supervised learning where only sparse pose annotations are needed, potentially alleviating the burden of meticulous mesh annotation. The effectiveness and robustness of our method are substantiated through both fully- and weakly-supervised experiments, demonstrating superior performance compared to state-of-the-art methods. We plan to release our code at https://github.com/Kun-Gao/P_GLFnet.

Self-Supervised 3D Semantic Occupancy Prediction from Multi-View 2D Surround Images

An accurate 3D representation of the geometry and semantics of an environment builds the basis for a large variety of downstream tasks and is essential for autonomous driving related tasks such as path planning and obstacle avoidance. The focus of this work is put on 3D semantic occupancy prediction, i.e., the reconstruction of a scene as a voxel grid where each voxel is assigned both an occupancy and a semantic label. We present a Convolutional Neural Network-based method that utilizes multiple color images from a surround-view setup with minimal overlap, together with the associated interior and exterior camera parameters as input, to reconstruct the observed environment as a 3D semantic occupancy map. To account for the ill-posed nature of reconstructing a 3D representation from monocular 2D images, the image information is integrated over time: Under the assumption that the camera setup is moving, images from consecutive time steps are used to form a multi-view stereo setup. In exhaustive experiments, we investigate the challenges presented by dynamic objects and the possibilities of training the proposed method with either 3D or 2D reference data. Latter being motivated by the comparably higher costs of generating and annotating 3D ground truth data. Moreover, we present and investigate a novel self-supervised training scheme that does not require any geometric reference data, but only relies on sparse semantic ground truth. An evaluation on the Occ3D dataset, including a comparison against current state-of-the-art self-supervised methods from the literature, demonstrates the potential of our self-supervised variant.

Digital 3D Hologram Generation Using Spatial and Elevation Information

The evolution of cartography poses challenges in representing three-dimensional terrain accurately on traditional two-dimensional maps. Providing an accurate 3D view of the area, coupled with essential geographic information, is vital for rapid and accurate decision-making in emergency management and response. Holography offers a promising solution by providing immersive three-dimensional visualizations. The field of hologram mapping, although novel, is still developing. Given its nascent stage, several limitations are evident. This study addresses one such limitation—inaccuracies in distance measurement—by presenting a hologram map that integrates two-dimensional and three-dimensional information. Accurate distance information on maps is critical for operational success. We aimed to improve hologram maps by integrating contour lines. Our approach allows users to measure distances from near-perpendicular angles while viewing 3D features from other perspectives. We review current advancements in hologram mapping, highlight existing limitations, and introduce our innovative solution designed to enhance both accuracy and usability. Our experiment resulted in a hologram map that accurately depicts a 3D environment, integrates contour lines, and allows for distance and slope angle measurements. The hologram map fills the research gap by providing accurate 3D visualization and distance measurement, signifying a major advancement in hologram mapping.

Analysis of the effect of geometric and flow parameters on the separation efficiency of virtual impact separators: Optimization based on response surface methodology

Respiratory dust is the primary factor triggering pneumoconiosis. Accurately separating respiratory dust from total dust according to specific criteria are key aspects of respiratory dust monitoring and prevention technology. In this study, we numerically solve the gas-solid two-phase flow within a virtual impactor by coupling computational fluid dynamics (CFD) with the discrete element method (DEM). Here, we explored the air velocity and dust separation efficiency of each component of the virtual impactor to determine optimal structural scale parameters. Accurate 3D printing technology was employed to materialize the three-dimensional structure of the virtual impactor. The separation efficiency of the virtual impactor was evaluated and verified through simulated duct experiments. The results indicate that the weak flow ratio has the most significant effect on separation efficiency, followed by the effect of the ratio of separation chamber diameter to nozzle width (S/W), the ratio of the weak flow outlet size to nozzle width (D/W) aving the smallest degree of influence. When S/W is 1.8, D/W is 1.333, and the weak flow ratio is 0.1, the separation efficiency curves show optimal performance, as confirmed by experimental verification. The experimental verification matched the simulated separation efficiency data within a deviation range of 1.10–11.4 %.

Optimizing Additive Manufacturable Structures with Computer Vision to Enhance Material Efficiency and Structural Stability

This study introduces an innovative technique that merges computer vision with topology optimization to advance additive manufacturing. Employing advanced photogrammetry software, we obtain high-resolution images of the design domain, which are then used to develop accurate 3D models through meticulous scanning procedures. These models are transformed into an STL file format and remeshed using an adaptive algorithm within COMSOL 5.3 Multiphysics, facilitated by a custom MATLAB 2023 application. This integration achieves the optimal mesh resolution and precision in analytical assessments. We applied this technique to the design of a concrete pillar for 3D printing, targeting a 75% reduction in volume to improve the material efficiency and structural stability—critical factors for extraterrestrial applications. The design, captured with a 360-degree camera array, guided the MATLAB-based topology optimization process. By combining MATLAB’s optimization algorithms with COMSOL’s meshing and finite element analysis tools, we investigated various material-efficient configurations. The findings reveal a substantial volume reduction, especially in the central region of the design, effectively optimizing material utilization while preserving structural integrity. The optimization algorithm exhibited a swift and stable convergence, reaching near-optimal solutions within approximately 20 iterations.

Interactions of Potential Endocrine-Disrupting Chemicals with Whole Human Proteome Predicted by AlphaFold2 Using an In Silico Approach.

Binding with proteins is a critical molecular initiating event through which environmental pollutants exert toxic effects in humans. Previous studies have been limited by the availability of three-dimensional (3D) protein structures and have focused on only a small set of environmental contaminants. Using the highly accurate 3D protein structure predicted by AlphaFold2, this study explored over 60 million interactions obtained through molecular docking between 20,503 human proteins and 1251 potential endocrine-disrupting chemicals. A total of 66,613,773 docking results were obtained, 1.2% of which were considered to be high binding, as their docking scores were lower than -7. Monocyte to macrophage differentiation factor 2 (MMD2) was predicted to interact with the highest number of environmental pollutants (526), with polychlorinated biphenyls and polychlorinated dibenzofurans accounting for a significant proportion. Dimension reduction and clustering analysis revealed distinct protein profiles characterized by high binding affinities for perfluoroalkyl and polyfluoroalkyl substances (PFAS), phthalate-like chemicals, and other pollutants, consistent with their uniquely enriched pathways. Further structural analysis indicated that binding pockets with a high proportion of charged amino acid residues, relatively low α-helix content, and high β-sheet content were more likely to bind to PFAS than others. This study provides insights into the toxicity pathways of various pollutants impacting human health and offers novel perspectives for the establishment and expansion of adverse outcome pathway-based models.

An efficient 3D tube-level flux-thermal model of external tubular receivers in solar power tower systems

Accurate and efficient radiative flux simulation and thermal analysis of liquid tubular receivers are essential for safe and stable operations of solar power tower systems. However, the complicated structure and photothermal processes on these receivers pose significant challenges for simulation. Although existing models have made great simplifications, the simulation results are still inefficient and inaccurate. To address these issues, a novel and universal 3D flux-thermal model is proposed. Firstly, a flux model is proposed via a tube-level Monte Carlo ray tracing method based on GPU parallel computing, enabling real-time and accurate radiative flux density distribution simulation for each absorber tube. Secondly, a high-resolution multi-tube thermal model is proposed, providing accurate 3D thermal analysis for each absorber tube. Through pre-factorization and highly parallelized designs, the proposed thermal model achieves a 600 times acceleration in efficiency, even with tens of millions of discrete elements. Finally, the novel flux-thermal model is validated against publicly published measured data from Solar Two, indicating a mean deviation of only 0.1% and a root mean square error of 0.006, demonstrating its high accuracy. Furthermore, the proposed flux-thermal model reveals complex and asymmetrical flux and temperature distributions on the absorber tube, in contrast to the symmetrical distributions obtained by the simplified models. Such inaccuracies in the simplified simulations may cause operational misjudgements and affect system safety potentially. Therefore, the proposed 3D tube-level flux-thermal model provides an accurate and efficient solution to radiative flux simulation and thermal analysis, enhancing the safe and stable operation of solar power tower systems.

DBMHT: A double-branch multi-hypothesis transformer for 3D human pose estimation in video

The estimation of 3D human poses from monocular videos presents a significant challenge. The existing methods face the problems of deep ambiguity and self-occlusion. To overcome these problems, we propose a Double-Branch Multi-Hypothesis Transformer (DBMHT). In detail, we utilize a Double-Branch architecture to capture temporal and spatial information and generate multiple hypotheses. To merge these hypotheses, we adopt a lightweight module to integrate spatial and temporal representations. The DBMHT can not only capture spatial information from each joint in the human body and temporal information from each frame in the video but also merge multiple hypotheses that have different spatio-temporal information. Comprehensive evaluation on two challenging datasets (i.e. Human3.6M and MPI-INF-3DHP) demonstrates the superior performance of DBMHT, marking it as a robust and efficient approach for accurate 3D HPE in dynamic scenarios. The results show that our model surpasses the state-of-the-art approach by 1.9% MPJPE with ground truth 2D keypoints as input.

AL-MobileNet: a novel model for 2D gesture recognition in intelligent cockpit based on multi-modal data

As the degree of automotive intelligence increases, gesture recognition is gaining more attention in human-vehicle interaction. However, existing gesture recognition methods are computationally intensive and perform poorly in multi-modal sensor scenarios. This paper proposes a novel network structure, AL-MobileNet (MobileNet with Attention and Lightweight Modules), which can quickly and accurately estimate 2D gestures in RGB and infrared (IR) images. The innovations of this paper are as follows: Firstly, to enhance multi-modal data, we created a synthetic IR dataset based on real 2D gestures and employed a coarse-to-fine training approach. Secondly, to speed up the model's computation on edge devices, we introduced a new lightweight computational module called the Split Channel Attention Block (SCAB). Thirdly, to ensure the model maintains accuracy in large datasets, we incorporated auxiliary networks and Angle-Weighted Loss (AWL) into the backbone network. Experiments show that AL-MobileNet requires only 0.4 GFLOPs of computational power and 1.2 million parameters. This makes it 1.5 times faster than MobileNet and allows for quick execution on edge devices. AL-MobileNet achieved a running speed of up to 28 FPS on the Ambarella CV28. On both general datasets and our dataset, our algorithm achieved an average PCK0.2 score of 0.95. This indicates that the algorithm can quickly generate accurate 2D gestures. The demonstration of the algorithm can be reviewed in gesturebaolong.

Exploring swine oviduct anatomy through micro-computed tomography: a 3D modeling perspective.

The oviduct plays a crucial role in the reproductive process, serving as the stage for fertilization and the early stages of embryonic development. When the environment of this organ has been mimicked, it has been shown to enhance in vitro embryo epigenetic reprogramming and to improve the yield of the system. This study explores the anatomical intricacies of two oviduct regions, the uterotubal junction (UTJ) and the ampullary-isthmic junction (AIJ) by using micro-computed tomography (MicroCT). In this study, we have characterized and 3D-reconstructed the oviduct structure, by measuring height and width of the oviduct's folds, along with the assessments of fractal dimension, lacunarity and shape factor. Results indicate distinct structural features in UTJ and AIJ, with UTJ displaying small, uniformly distributed folds and high lacunarity, while AIJ shows larger folds with lower lacunarity. Fractal dimension analysis reveals values for UTJ within 1.189-1.1779, while AIJ values range from 1.559-1.770, indicating differences in structural complexity between these regions. Additionally, blind sacs or crypts are observed, akin to those found in various species, suggesting potential roles in sperm sequestration or reservoir formation. These morphological differences align with functional variations and are essential for developing an accurate 3D model. In conclusion, this research provides information about the oviduct anatomy, leveraging MicroCT technology for detailed 3D reconstructions, which can significantly contribute to the understanding of geometric-morphological characteristics influencing functional traits, providing a foundation for a biomimetic oviduct-on-a-chip.

Driving mutual advancement of 3D reconstruction and inpainting for masked faces

Target occlusion or pollution has always been a common and difficult problem in 3D reconstruction, seriously affecting the reconstruction effect, especially in single image scenario. To address the issues of incomplete reconstruction caused by pixel missing, a novel framework promoting mutual enhancement between 3D reconstruction and inpainting is proposed, which is capable of reconstructing a realistic 3D face with pixel completion from single image without prior assumptions. The framework is composed of an inpainting module for facial pixel completion and a reconstruction module for 3D face modeling. The inpainting module provides complete texture and accurate 3D geometry inferring premise for 3D reconstruction, and the reconstruction module is also involved in face inpainting, constraining the inpainting module to fill plausible pixels that are closer to the ground truth. Experiments show that the accurate 3D faces with complete and fine texture can be reconstructed by the proposed framework from largely masked images, with a competitive performance even surpassing most of state-of-the-art methods with unmasked images as input. Furthermore, this framework has achieved state-of-the-art performance in image inpainting on multiple face datasets.

UNETR++: Delving Into Efficient and Accurate 3D Medical Image Segmentation.

Owing to the success of transformer models, recent works study their applicability in 3D medical segmentation tasks. Within the transformer models, the self-attention mechanism is one of the main building blocks that strives to capture long-range dependencies, compared to the local convolutional-based design. However, the self-attention operation has quadratic complexity which proves to be a computational bottleneck, especially in volumetric medical imaging, where the inputs are 3D with numerous slices. In this paper, we propose a 3D medical image segmentation approach, named UNETR++, that offers both high-quality segmentation masks as well as efficiency in terms of parameters, compute cost, and inference speed. The core of our design is the introduction of a novel efficient paired attention (EPA) block that efficiently learns spatial and channel-wise discriminative features using a pair of inter-dependent branches based on spatial and channel attention. Our spatial attention formulation is efficient and has linear complexity with respect to the input. To enable communication between spatial and channel-focused branches, we share the weights of query and key mapping functions that provide a complimentary benefit (paired attention), while also reducing the complexity. Our extensive evaluations on five benchmarks, Synapse, BTCV, ACDC, BraTS, and Decathlon-Lung, reveal the effectiveness of our contributions in terms of both efficiency and accuracy. On Synapse, our UNETR++ sets a new state-of-the-art with a Dice Score of 87.2%, while significantly reducing parameters and FLOPs by over 71%, compared to the best method in the literature. Our code and models are available at: https://tinyurl.com/2p87x5xn.

Effect of twin-screw extrusion pretreatment on starch structure, rheological properties and 3D printing accuracy of whole potato flour and its application in dysphagia diets

Twin-screw extrusion pretreatment has great potential for the development of three-dimensional (3D) printed food as dysphagia diets. This study aimed to investigate the effect of twin-screw extrusion pretreatment on starch structure, rheological properties and 3D printing accuracy of whole potato flour and its application in dysphagia diets. The results indicated that twin-screw extrusion pretreatment was found to change chain length distributions, short-range ordered structure and relative crystallinity of whole potato flour (WPF), thereby improving its 3D printing performance. With the increasing proportion of long linear chains (DP > 12), the intensity of hydrogen bonds, linear viscoelastic region, storage modulus (G′), loss modulus (G″), viscosity and n of whole potato flour paste were increased, enhancing high printing accuracy and shape retention of 3D printed samples with a denser microstructure and smaller pore diameter distribution. The whole potato flour paste extruded with a peristaltic pump speed at 5.25 mL/min (WPF-4) displayed the highest printing accuracy with excellent rheological properties, good water distribution state and dense network structure, which classified as class 5 level dysphagia diets. This research provides an effective guidance for the modification of whole potato flour using twin-screw extrusion pretreatment as 3D printed food inks for dysphagia patients.

DIPHORM: An Innovative DIgital PHOtogrammetRic Monitoring Technique for Detecting Surficial Displacements of Landslides

Monitoring surface displacements of landslides is essential for evaluating their evolution and the effectiveness of mitigation works. Traditional methods like robotic total stations (RTSs) and GNSS provide high-accuracy measurements but are limited to discrete points, potentially missing the broader landslide’s behavior. On the contrary, laser scanner surveys offer accurate 3D representations of slopes and the possibility of inferring their movements, but they are often limited to infrequent, high-cost surveys. Monitoring techniques based on ground-based digital photogrammetry may represent a new, robust, and cost-effective alternative. This study demonstrates the use of multi-temporal images from fixed and calibrated cameras to achieve the 3D reconstruction of landslide displacements. The method presented offers the important benefit of obtaining spatially dense displacement data across the entire camera view and quasi-continuous temporal measurement. This paper outlines the framework for this prototyping technique, along with a description of the necessary hardware and procedural steps. Furthermore, strengths and weaknesses are discussed based on the activities carried out in a landslide case study in northeastern Italy. The results from the photo-monitoring are reported, discussed, and compared with traditional topographical data, validating the reliability of this new approach in monitoring the time evolution of surface displacements across the entire landslide area.

Deep Learning-Based Workflow for Bone Segmentation and 3D Modeling in Cone-Beam CT Orthopedic Imaging

In this study, a deep learning-based workflow designed for the segmentation and 3D modeling of bones in cone beam computed tomography (CBCT) orthopedic imaging is presented. This workflow uses a convolutional neural network (CNN), specifically a U-Net architecture, to perform precise bone segmentation even in challenging anatomical regions such as limbs, joints, and extremities, where bone boundaries are less distinct and densities are highly variable. The effectiveness of the proposed workflow was evaluated by comparing the generated 3D models against those obtained through other segmentation methods, including SegNet, binary thresholding, and graph cut algorithms. The accuracy of these models was quantitatively assessed using the Jaccard index, the Dice coefficient, and the Hausdorff distance metrics. The results indicate that the U-Net-based segmentation consistently outperforms other techniques, producing more accurate and reliable 3D bone models. The user interface developed for this workflow facilitates intuitive visualization and manipulation of the 3D models, enhancing the usability and effectiveness of the segmentation process in both clinical and research settings. The findings suggest that the proposed deep learning-based workflow holds significant potential for improving the accuracy of bone segmentation and the quality of 3D models derived from CBCT scans, contributing to better diagnostic and pre-surgical planning outcomes in orthopedic practice.

Three-Dimensional Posture Estimation of Vehicle Occupants Using Depth and Infrared Images.

Pose estimation is crucial for ensuring passenger safety and better user experiences in semi- and fully autonomous vehicles. Traditional methods relying on pose estimation from regular color images face significant challenges due to a lack of three-dimensional (3D) information and the sensitivity to occlusion and lighting conditions. Depth images, which are invariant to lighting issues and provide 3D information about the scene, offer a promising alternative. However, there is a lack of strong work in 3D pose estimation from such images due to the time-consuming process of annotating depth images with 3D postures. In this paper, we present a novel approach to 3D human posture estimation using depth and infrared (IR) images. Our method leverages a three-stage fine-tuning process involving simulation data, approximated data, and a limited set of manually annotated samples. This approach allows us to effectively train a model capable of accurate 3D pose estimation with a median error of under 10 cm across all joints, using fewer than 100 manually annotated samples. To the best of our knowledge, this is the first work focusing on vehicle occupant posture detection utilizing only depth and IR data. Our results demonstrate the feasibility and efficacy of this approach, paving the way for enhanced passenger safety in autonomous vehicle systems.