3D Reconstruction Research Articles

Study on the impact of camera ISP algorithms on stripe structured light 3D reconstruction and the improvement of 3D reconstruction accuracy by photon input

This paper investigates the impact of camera image signal processing (ISP) algorithms on stripe structured light 3D reconstruction and explores the improvement of 3D reconstruction accuracy by photon input. Existing 3D reconstruction technologies have broad applications in fields such as intelligent manufacturing, healthcare, and consumer electronics, but their high precision requirements often cannot be met by current ISP algorithms. By using handheld devices to capture images and combining key techniques such as the multi-step phase shifting method, multi-frequency phase unwrapping method, and triangulation method, this paper conducts an in-depth study on the application of photon input in 3D reconstruction. The study demonstrates that using non-visual information (RAW images) as input can significantly improve reconstruction accuracy, producing more accurate results compared to images processed by ISP. The paper also quantifies the impact of ISP processing on 3D reconstruction results by comparing the depth information and point cloud data of two sets of images. Experimental results show that disabling certain ISP algorithms, such as bilateral noise filtering (BNF), edge enhancement (EEH), and non-local means denoising (NLM), can further improve reconstruction accuracy and reduce errors. In conclusion, this paper proposes a photon image-based 3D reconstruction method that, combined with artificial intelligence technology and differentiable point cloud rendering techniques, holds promise for achieving higher precision and faster 3D reconstruction. This technology is of great significance in practical applications, particularly in the field of industrial close-range 3D reconstruction.

3DRecNet: A 3D Reconstruction Network with Dual Attention and Human-Inspired Memory

Humans inherently perceive 3D scenes using prior knowledge and visual perception, but 3D reconstruction in computer graphics is challenging due to complex object geometries, noisy backgrounds, and occlusions, leading to high time and space complexity. To addresses these challenges, this study introduces 3DRecNet, a compact 3D reconstruction architecture optimized for both efficiency and accuracy through five key modules. The first module, the Human-Inspired Memory Network (HIMNet), is designed for initial point cloud estimation, assisting in identifying and localizing objects in occluded and complex regions while preserving critical spatial information. Next, separate image and 3D encoders perform feature extraction from input images and initial point clouds. These features are combined using a dual attention-based feature fusion module, which emphasizes features from the image branch over those from the 3D encoding branch. This approach ensures independence from proposals at inference time and filters out irrelevant information, leading to more accurate and detailed reconstructions. Finally, a Decoder Branch transforms the fused features into a 3D representation. The integration of attention-based fusion with the memory network in 3DRecNet significantly enhances the overall reconstruction process. Experimental results on the benchmark datasets, such as ShapeNet, ObjectNet3D, and Pix3D, demonstrate that 3DRecNet outperforms existing methods.

The larval chondrocranium and its development in Smilisca phaeota with considerations of patterns characteristic for the chondrocranial development of Lalagobatrachia

Several studies describe the development of the chondrocranium of vertebrates. The details in these studies vary a lot, which makes it hard to compare developmental patterns and identify evolutionary trends. Therefore, we aim to close this gap for anurans, which is the largest order of amphibians. We present here a detailed description of the chondrocranium morphology and development of Smilisca phaeota, the New Granada cross-banded tree frog. The anatomy was described for the larvae at or older than Gossner stage 31 and before ossification starts. Following this, we describe the development of the chondrocranium from Gossner stages 19–26. Early in Gossner stage 19 no precursors of any cartilages are visible, while later in that stage the mesodermal Anlage of Meckel’s cartilage was observed. In the subsequent stages more and more mesodermal anlagen become identifiable, followed by chondrification, and final differentiation of the cartilage elements. We used serial sections to study all the developmental stages and additionally utilized cleared and stained specimens and CT scan data. The latter were also used for the 3D reconstruction of the chondrocranium. We previously studied several species and compared these developmental patterns with S. phaeota, revealing potentially characteristic patterns significant for Lalagobatrachia, a clade that includes over 7000 frog species. These include (1) the suprarostral alae develop before the suprarostral corpus, (2) the infrarostral cartilage chondrifies late, after the chondrification of ceratobranchial 1, and (3) the ceratohyal body is the first element to show chondrocytes and to chondrify. However, with only six species studied so far, our data only provide a basis for future studies and developing hypotheses about the ancestral developmental pattern in anurans.

A new insight for investigating the prenatal and postnatal ossification centers of pelvic and femur bones in white New Zealand rabbits (Oryctolagus cuniculus) using 3D CT, double stain technique, and morphometry

BackgroundThe ossification centers in rabbit limbs are related to fetal age and bone maturation. Objective: To address the limited studies on ossification in the hind limbs of New Zealand rabbits, we investigated the prenatal and postnatal development of the pelvic and femur bones. MethodsDouble staining with Alcian Blue and Alizarin Red, computed tomography (CT), and 3D reconstruction were employed to visualize and analyze ossification centers in detail. ResultsUsing double staining, we observed these patterns: At prenatal days 18 and 21, ossification centers appeared in the ilium. By prenatal days 23 and 25, ossification began in the ischium. On postnatal day 1, ilium ossification centers spread across most of the ilium wings, except for the iliac crest, and new centers appeared in the pubis and cotyloid bones. Most bones had ossified by the third week and one month postnatal, except for the iliac crest and ischial tuberosity. At 1.5 months, both were fully ossified. On day 18 post coitum, an ossification center was visible in the middle of the femur shaft. By day 28 post coitum, ossification extended through the shaft, and postnatally, new ossification spots appeared at the extremities by day one and week one. By the third week, complete ossification of the femur head, lesser trochanter, third trochanter, medial condyle, and lateral condyle was observed. At 1.5 months, the entire proximal extremity was ossified. Conclusion3D CT provided clear imaging of ossification progression in the pelvic and femur bones. This study enhances our understanding of vertebrate skeletal development.

Automated generation and semantic segmentation of roof orthophoto for digital twin -based monitoring of slated roofs

Roofs are one of the building elements most exposed to environment-induced deterioration, and their deterioration can rapidly result in potential safety hazards to occupants or even buildings’ failure. In pursuit of cost-effective, efficient and pro-active building roof monitoring, this paper presents a computing pipeline for: 1. Automated roof orthophoto generation that uses as input the 3D reconstruction from structure-from-motion photogrammetry and the building’s Digital Twin (DT) model, and outputs occlusion-free panel orthophotos with metric resolution. 2. Automated orthophoto semantic segmentation using deep learning to extract visible roof sub-components from the orthophotos. Various segmentation models are tested and contrasted. The approaches are validated with three real case studies of traditional buildings with slated pitched roofs. The proposed approach employs a generic pipeline which could be applied to any element of the building fabric.

A domain adaptation framework for cross-modality SAR 3D reconstruction point clouds segmentation utilizing LiDAR data

Synthetic aperture radar (SAR) 3D point clouds reconstruction can eliminate the problems of layover in 2D SAR image projections, the recognition of the reconstructed point cloud can significantly enhance target identification and information extraction. Current SAR 3D point cloud segmentation methods, based on traditional machine learning clustering approaches, suffer from poor automation and accuracy. However, the absence of publicly labeled SAR datasets impedes the progress of deep learning-based SAR 3D point cloud segmentation methods. To tackle the aforementioned challenges, we introduce, for the first time, an alternative training approach for SAR 3D point cloud segmentation using LiDAR annotated data, which offers a more abundant sample pool, thus alleviating the lack of training sets. Nevertheless, a segmentation model trained on LiDAR point clouds exhibits a significant decline in performance when directly applied to SAR 3D reconstructed point clouds due to the cross-domain discrepancy. This research presents a pioneering domain adaptation 3D semantic segmentation framework to implement cross-modal learning for SAR point clouds. In our scheme, a simple yet effective technique is developed to achieve segmentation of SAR double-bounce scattering regions called SARDBS-Mix, which employs a mixing strategy, to capture the distinctive reflection characteristics of SAR data. Furthermore, we implement approaches including center alignment and normalization, local augmentation, and weighted cross entropy to mitigate LiDAR and SAR domain gap and class imbalances. The experimental results validate the feasibility and effectiveness of the proposed method for SAR 3D reconstructed point cloud segmentation.

A Study on the 3D Reconstruction Strategy of a Sheep Body Based on a Kinect v2 Depth Camera Array.

Non-contact measurement based on the 3D reconstruction of sheep bodies can alleviate the stress response in sheep during manual measurement of body dimensions. However, data collection is easily affected by environmental factors and noise, which is not conducive to practical production needs. To address this issue, this study proposes a non-contact data acquisition system and a 3D point cloud reconstruction method for sheep bodies. The collected sheep body data can provide reference data for sheep breeding and fattening. The acquisition system consists of a Kinect v2 depth camera group, a sheep passage, and a restraining pen, synchronously collecting data from three perspectives. The 3D point cloud reconstruction method for sheep bodies is implemented based on C++ language and the Point Cloud Library (PCL). It processes noise through pass-through filtering, statistical filtering, and random sample consensus (RANSAC). A conditional voxel filtering box is proposed to downsample and simplify the point cloud data. Combined with the RANSAC and Iterative Closest Point (ICP) algorithms, coarse and fine registration are performed to improve registration accuracy and robustness, achieving 3D reconstruction of sheep bodies. In the base, 135 sets of point cloud data were collected from 20 sheep. After 3D reconstruction, the reconstruction error of body length compared to the actual values was 0.79%, indicating that this method can provide reliable reference data for 3D point cloud reconstruction research of sheep bodies.

Automated Measurement of Cattle Dimensions Using Improved Keypoint Detection Combined with Unilateral Depth Imaging.

Traditional measurement methods often rely on manual operations, which are not only inefficient but also cause stress to cattle, affecting animal welfare. Currently, non-contact cattle dimension measurement usually involves the use of multi-view images combined with point cloud or 3D reconstruction technologies, which are costly and less flexible in actual farming environments. To address this, this study proposes an automated cattle dimension measurement method based on an improved keypoint detection model combined with unilateral depth imaging. Firstly, YOLOv8-Pose is selected as the keypoint detection model and SimSPPF replaces the original SPPF to optimize spatial pyramid pooling, reducing computational complexity. The CARAFE architecture, which enhances upsampling content-aware capabilities, is introduced at the neck. The improved YOLOv8-pose achieves a mAP of 94.4%, a 2% increase over the baseline model. Then, cattle keypoints are captured on RGB images and mapped to depth images, where keypoints are optimized using conditional filtering on the depth image. Finally, cattle dimension parameters are calculated using the cattle keypoints combined with Euclidean distance, the Moving Least Squares (MLS) method, Radial Basis Functions (RBFs), and Cubic B-Spline Interpolation (CB-SI). The average relative errors for the body height, lumbar height, body length, and chest girth of the 23 measured beef cattle were 1.28%, 3.02%, 6.47%, and 4.43%, respectively. The results show that the method proposed in this study has high accuracy and can provide a new approach to non-contact beef cattle dimension measurement.

Domain adaptation strategies for 3D reconstruction of the lumbar spine using real fluoroscopy data

In this study, we address critical barriers hindering the widespread adoption of surgical navigation in orthopedic surgeries due to limitations such as time constraints, cost implications, radiation concerns, and integration within the surgical workflow. Recently, our work X23D showed an approach for generating 3D anatomical models of the spine from only a few intraoperative fluoroscopic images. This approach negates the need for conventional registration-based surgical navigation by creating a direct intraoperative 3D reconstruction of the anatomy. Despite these strides, the practical application of X23D has been limited by a significant domain gap between synthetic training data and real intraoperative images.In response, we devised a novel data collection protocol to assemble a paired dataset consisting of synthetic and real fluoroscopic images captured from identical perspectives. Leveraging this unique dataset, we refined our deep learning model through transfer learning, effectively bridging the domain gap between synthetic and real X-ray data. We introduce an innovative approach combining style transfer with the curated paired dataset. This method transforms real X-ray images into the synthetic domain, enabling the in-silico-trained X23D model to achieve high accuracy in real-world settings.Our results demonstrated that the refined model can rapidly generate accurate 3D reconstructions of the entire lumbar spine from as few as three intraoperative fluoroscopic shots. The enhanced model reached a sufficient accuracy, achieving an 84% F1 score, equating to the benchmark set solely by synthetic data in previous research. Moreover, with an impressive computational time of just 81.1 ms, our approach offers real-time capabilities, vital for successful integration into active surgical procedures.By investigating optimal imaging setups and view angle dependencies, we have further validated the practicality and reliability of our system in a clinical environment. Our research represents a promising advancement in intraoperative 3D reconstruction. This innovation has the potential to enhance intraoperative surgical planning, navigation, and surgical robotics.

A novel 3D reconstruction method of blast furnace burden surface based on virtual camera array

Accurate three-dimensional (3D) blast furnace burden surface topography is crucial for optimizing material distribution and furnace operation. However, harsh conditions within the furnace pose challenges to this 3D reconstruction, such as difficulties in matching feature points and sparse 3D point clouds. To overcome these challenges, we propose a 3D reconstruction method for the blast furnace burden surface using virtual camera array. Firstly, an ORB feature extraction and matching method based on bidirectional optical flow tracking is proposed to solve the impact of illumination changes and noise. Then, the virtual camera array is constructed to generate a sparse 3D point cloud from an image sequence. Finally, dense surface reconstruction of the burden surface is achieved based on the improved PMVS and the Poisson surface reconstruction method. Experimental results demonstrate that the proposed method can extract accurate feature points and reconstruct the accurate 3D topography of the burden surface.

OpenECAD: An efficient visual language model for editable 3D-CAD design

Computer-aided design (CAD) tools are utilized in the manufacturing industry for modeling everything from cups to spacecraft. These programs are complex to use and typically require years of training and experience to master. Structured and well-constrained 2D sketches and 3D constructions are crucial components of CAD modeling. A well-executed CAD model can be seamlessly integrated into the manufacturing process, thereby enhancing production efficiency. Deep generative models of 3D shapes and 3D object reconstruction models have garnered significant research interest. However, most of these models produce discrete forms of 3D objects that are not editable. Moreover, the few models based on CAD operations often have substantial input restrictions. In this work, we fine-tuned pre-trained models to create OpenECAD models (0.55B, 0.89B, 2.4B and 3.1B), leveraging the visual, logical, coding, and general capabilities of visual language models. OpenECAD models can process images of 3D designs as input and generate highly structured 2D sketches and 3D construction commands, ensuring that the designs are editable. These outputs can be directly used with existing CAD tools’ APIs to generate project files. To train our network, we created a series of OpenECAD datasets. These datasets are derived from existing public CAD datasets, adjusted and augmented to meet the specific requirements of vision language model (VLM) training. Additionally, we have introduced an approach that utilizes dependency relationships to define and generate sketches, further enriching the content and functionality of the datasets.

Microscopic interface deterioration mechanism and damage behavior of high-toughness recycled aggregate concrete based on 4D in-situ CT experiments

High-toughness recycled aggregate concrete (HTRAC), as an innovative green building material, showcases outstanding performance and environmental attributes. To investigate its microscopic structural characteristics and the mechanisms of damage evolution under loading, a series of 4D in-situ X-ray computed tomography (CT) experiments were conducted under uniaxial compression conditions. During the experimental process, detailed scans were performed on six key loading points (0 %, 30 %, 60 %, and 100 % of the peak load, 70 % and 40 % of the post-peak load of the specimen). Advanced three-dimensional image processing techniques were utilized to conduct a comprehensive analysis of internal features within the specimen, including pores, steel fibers, and cracks. The evolution of damage in HTRAC was characterized by both qualitative 2D crack slice analysis and quantitative 3D reconstruction analysis of cracks. Changes in the average grayscale values of CT cross-sectional images were also examined. The deterioration mechanism of recycled aggregate concrete interface cracks was elucidated by the combination of CT images and modeled HTRAC. Furthermore, a statistics model illustrating the interaction mechanism between damage, cracks, and strain was proposed. This study provides technical and theoretical support for the further promotion and application of high-toughness recycled aggregate concrete.

Identification and mechanism of wheat protein disulfide isomerase-promoted gluten network formation.

Formation of the gluten network depends on glutenin crosslinking via disulfide bonds, and wheat protein disulfide isomerase (wPDI) plays an important role in this process. Here, we identify a substrate gluten protein of wPDI and the mechanism underlying wPDI-promoted glutenin crosslinking. Farinographic, rheologic, and alveographic analysis unambiguously proves that wPDI improves gluten network formation, which is directly observed by 3D reconstruction of the gluten network. Protein analysis and LC-MS/MS reveal that glutenin subunit 1Dx5 is primarily recruited by wPDI to participate in gluten network formation, and its cysteine-containing N-terminal domain (1Dx5-NTD), which harbors three cysteine residues for crosslinking, is purified. 1Dx5-NTD interacts with wPDI in both redox states, possibly folded by reduced wPDI and then catalyzed by oxidized wPDI, as further evidenced by wPDI-promoted self-crosslinking. Consistent with macroscopic observations, our results suggest that wPDI folds 1Dx5-NTD into β-strand structure that favors disulfide bond formation.

Mapping distribution of fractures and minerals in rock samples using Res-VGG-UNet and threshold segmentation methods

Understanding the internal structures of rock and their subsequent evolution is essential across multiple disciplines. Fractures and mineral grains are significant components of internal structures. These components influence natural processes and human activities, such as oil migration and extraction. With advancements in non-destructive evaluation (NDE) techniques, especially X-ray computed tomography (CT), in-depth analyses of internal structures of rock have become feasible. However, the data-intensive nature of CT scans necessitates more efficient post-processing algorithms. This research uses the Residual Network-Visual Geometry Group-UNet (Res-VGG-UNet) deep learning framework in conjunction with two threshold segmentation methods to capture components like fractures, the drilled hole, and minerals. The performance of Res-VGG-UNet is evaluated against DeepLabv3+ and Weka3D. Key findings include the capability of the deep learning technique to distinguish between various components and the performance of single-slice-based versus multi-slice-based threshold segmentation methods. VGG16 and VGG19 outperform DeepLabv3+ and Weka3D in the performance of fracture classification. Moreover, a 3D reconstruction technique is introduced, providing enhanced insights into internal features of the rock. We conclude this paper by showing the potential to use the outcomes of the deep learning and multi-slice-based threshold segmentation methods to develop discrete element models, accurately reflecting the physical characteristics of rock specimens.

3D modelling method and application to a digital campus by fusing point cloud data and image data

ObjectiveThe use of single-source data for real-world 3D modelling currently faces problems such as deformation, pulling and fuzzy texture at the bottom of buildings in some feature models because of the lack of images. Moreover, LIDAR generates a huge amount of data, and the massive raw data processing and point cloud parsing puts high demands on the hardware arithmetic and algorithms. Aiming at the deficiencies and defects of the two data sources of inclined photogrammetry and airborne laser point cloud in the construction of high-quality and high-precision city-level 3D models. Methodsthis study uses a university library building as an example and proposes the main technical process and method of modelling after fusing the point cloud data acquired by inclined photogrammetry and 3D laser scanning technology. This is accomplished in the reconstruction stage of multi-source data fusion through data spatial alignment, coordinate system unification and data spatial integration. At the stage of multi-source data fusion and reconstruction, through data spatial alignment, coordinate system unification, point cloud coarse alignment and the iterative closest point (ICP) algorithm, a realistic 3D model of a building is constructed to verify the effectiveness of the modelling method. ResultsThe method can effectively improve the accuracy of the real-life 3D model, repair the deficiencies in the model and optimise the details of the model. It can also significantly improve the fineness of the tilt photography model and perfectly present the geometric and texture information of the building, making it a superior method for fine 3D reconstruction. ConclusionThis 3D reconstruction method of buildings, which integrates low-altitude inclined photogrammetry and airborne light detection and ranging (LiDAR), has high positional accuracy and can provide new methods and new ideas for the construction of digital campuses as well as for other engineering applications.

An Improved 3D Reconstruction Method for Satellite Images Based on Generative Adversarial Network Image Enhancement

Three-dimensional reconstruction based on optical satellite images has always been a research hotspot in the field of photogrammetry. In particular, the 3D reconstruction of building areas has provided great help for urban planning, change detection and emergency response. The results of 3D reconstruction of satellite images are greatly affected by the input images, and this paper proposes an improvement method for 3D reconstruction of satellite images based on the generative adversarial network (GAN) image enhancement. In this method, the perceptual loss function is used to optimize the network, so that it can output high-definition satellite images for 3D reconstruction, so as to improve the completeness and accuracy of the reconstructed 3D model. We use the public benchmark dataset of satellite images to test the feasibility and effectiveness of the proposed method. The experiments show that compared with the satellite stereo pipeline (S2P) method and the bundle adjustment (BA) method, the proposed method can automatically reconstruct high-quality 3D point clouds.

A morphology-Euclidean-linear recognition method for rebar point clouds of highway tunnel linings during the construction phase

PurposeLaser point clouds are a 3D reconstruction method with wide range, high accuracy and strong adaptability. Therefore, the purpose is to discover a construction point cloud extraction method that can obtain complete information about the construction of rebar, facilitating construction quality inspection and tunnel data archiving, to reduce the cost and complexity of construction management.Design/methodology/approachFirstly, this paper analyzes the point cloud data of the tunnel during the construction phase, extracts the main features of the rebar data and proposes an M-E-L recognition method. Secondly, based on the actual conditions of the tunnel and the specifications of Chinese tunnel engineering, a rebar model experiment is designed to obtain experimental data. Finally, the feasibility and accuracy of the M-E-L recognition method are analyzed and tested based on the experimental data from the model.FindingsBased on tunnel morphology characteristics, data preprocessing, Euclidean clustering and PCA shape extraction methods, a M-E-L identification algorithm is proposed for identifying secondary lining rebars in highway tunnel construction stages. The algorithm achieves 100% extraction of the first-layer rebars, allowing for the three-dimensional visualization of the on-site rebar situation. Subsequently, through data processing, rebar dimensions and spacings can be obtained. For the second-layer rebars, 55% extraction is achieved, providing information on the rebar skeleton and partial rebar details at the construction site. These extracted data can be further processed to verify compliance with construction requirements.Originality/valueThis paper introduces a laser point cloud method for double-layer rebar identification in tunnels. Current methods rely heavily on manual detection, lacking objectivity. Objective approaches for automatic rebar identification include image-based and LiDAR-based methods. Image-based methods are constrained by tunnel lighting conditions, while LiDAR focuses on straight rebar skeletons. Our research proposes a 3D point cloud recognition algorithm for tunnel lining rebar. This method can extract double-layer rebars and obtain construction rebar dimensions, enhancing management efficiency.

Building Better Models: Benchmarking Feature Extraction and Matching for Structure from Motion at Construction Sites

The popularity of Structure from Motion (SfM) techniques has significantly advanced 3D reconstruction in various domains, including construction site mapping. Central to SfM, is the feature extraction and matching process, which identifies and correlates keypoints across images. Previous benchmarks have assessed traditional and learning-based methods for these tasks but have not specifically focused on construction sites, often evaluating isolated components of the SfM pipeline. This study provides a comprehensive evaluation of traditional methods (e.g., SIFT, AKAZE, ORB) and learning-based methods (e.g., D2-Net, DISK, R2D2, SuperPoint, SOSNet) within the SfM pipeline for construction site mapping. It also compares matching techniques, including SuperGlue and LightGlue, against traditional approaches such as nearest neighbor. Our findings demonstrate that deep learning-based methods such as DISK with LightGlue and SuperPoint with various matchers consistently outperform traditional methods like SIFT in both reconstruction quality and computational efficiency. Overall, the deep learning methods exhibited better adaptability to complex construction environments, leveraging modern hardware effectively, highlighting their potential for large-scale and real-time applications in construction site mapping. This benchmark aims to assist researchers in selecting the optimal combination of feature extraction and matching methods for SfM applications at construction sites.

3D point cloud reconstruction for array GM-APD lidar based on echo waveform decomposition

Array GM-APD lidar only reconstructs one response distance per pixel, resulting in 3D point clouds sparse. This paper proposes a 3D point cloud reconstruction algorithm using echo waveforms directly. First, combining empirical mode decomposition and the array GM-APD lidar echo waveform characteristics to extract multiple target response distances per pixel. Then, assuming per-pixel target surface rotation to resolve the shared azimuth and polar angles of multiple response distances. Calculate the rotation angles using the decomposed distance difference and by minimizing the plane fitting error. Finally, reconstruct the 3D point cloud by rotation translation transformation. Simulations and experiments demonstrate that the proposed algorithm reduces the Chamfer Distance of the reconstructed 3D point cloud by over 10 % compared to the original. The reconstructed 3D point cloud of buildings and vehicles shows lower sparsity and stepwise, and the reconstructed 3D point cloud for each target in the multiple-target scene is unaffected by others.

Label-free spatiotemporal decoding of single-cell fate via acoustic driven 3D tomography

Label-free three-dimensional imaging plays a crucial role in unraveling the complexities of cellular functions and interactions in biomedical research. Conventional single-cell optical tomography techniques offer affordability and the convenience of bypassing laborious cell labelling protocols. However, these methods are encumbered by restricted illumination scanning ranges on abaxial plane, resulting in the loss of intricate cellular imaging details. The ability to fully control cellular rotation across all angles has emerged as an optimal solution for capturing comprehensive structural details of cells. Here, we introduce a label-free, cost-effective, and readily fabricated contactless acoustic-induced vibration system, specifically designed to enable multi-degree-of-freedom rotation of cells, ultimately attaining stable in-situ rotation. Furthermore, by integrating this system with advanced deep learning technologies, we perform 3D reconstruction and morphological analysis on diverse cell types, thus validating groups of high-precision cell identification. Notably, long-term observation of cells reveals distinct features associated with drug-induced apoptosis in both cancerous and normal cells populations. This methodology, based on deep learning-enabled cell 3D reconstruction, charts a novel trajectory for groups of real-time cellular visualization, offering promising advancements in the realms of drug screening and post-single-cell analysis, thereby addressing potential clinical requisites.