3D Shape Model Research Articles

A 3D Statistical Shape Model of the Right Ventricular Outflow Tract in Pulmonary Valve Replacement Patients Post-Surgical Repair.

Assessment of the right ventricular outflow tract and pulmonary arteries (RVOT) for percutaneous pulmonary valve implantation (PPVI) uses discrete measurements (diameters and lengths) from medical images. This multi-centre study identified the 3D RVOT shape features prevalent in patients late after surgical repair of congenital heart disease (CHD). A 3D RVOT statistical shape model (SSM) was computed from 81 retrospectively selected CHD patients (14.7 ± 6.8 years) who required pulmonary valve replacement late after surgical repair. A principal component analysis identified prevalent shape features (modes) within the population which were compared with standard geometric measurements (diameter, length and surface area) and between sub-groups of diagnosis, RVOT type and dysfunction. Shape mode 1 and 2 represented RVOT size and curvature and tapering and length, respectively. Shape modes 3-5 related to branch pulmonary artery calibre, conical vs. bulbous RVOTs and RVOT curvature, respectively. Tetralogy of Fallot, transannular patch type and regurgitant RVOTs were larger and straighter while conduit and stenotic types were longer and more cylindrical than other subgroups. This SSM analysed the main 3D shape features present in a population of RVOTs, exploiting the wide 3D anatomical information provided by routine imaging. This morphological information may have implications for PPVI patient selection and device design.

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.

Leveraging 3D Atrial Geometry for the Evaluation of Atrial Fibrillation: A Comprehensive Review.

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia associated with significant morbidity and mortality. Managing risk of stroke and AF burden are pillars of AF management. Atrial geometry has long been recognized as a useful measure in achieving these goals. However, traditional diagnostic approaches often overlook the complex spatial dynamics of the atria. This review explores the emerging role of three-dimensional (3D) atrial geometry in the evaluation and management of AF. Advancements in imaging technologies and computational modeling have enabled detailed reconstructions of atrial anatomy, providing insights into the pathophysiology of AF that were previously unattainable. We examine current methodologies for interpreting 3D atrial data, including qualitative, basic quantitative, global quantitative, and statistical shape modeling approaches. We discuss their integration into clinical practice, highlighting potential benefits such as personalized treatment strategies, improved outcome prediction, and informed treatment approaches. Additionally, we discuss the challenges and limitations associated with current approaches, including technical constraints and variable interpretations, and propose future directions for research and clinical applications. This comprehensive review underscores the transformative potential of leveraging 3D atrial geometry in the evaluation and management of AF, advocating for its broader adoption in clinical practice.

Facial Masculinization Surgery Using Polyetheretherketone Alloplasty: Statistical Shape Modeling-based Implant Designs.

The face is the initial feature used to judge gender in public spaces; it is also a source of significant gender dysphoria. Surgical techniques are available for non-cisgender male patients who desire a more masculine face by augmenting certain features to change the bony framework of the skull. Augmentation using virtually designed patient-specific polyetheretherketone implants has now become a more widely applied method in maxillofacial surgery. When designing implants for augmentation, a three-dimensional (3D) reference or template is very useful. Hence, a 3D statistical shape model was developed of a male skull shape from information from a population of 40 male patients containing the mean shape and principal components of shape variation. By overlaying the template and the patient's 3D skull model, this method identified the regions of gender dimorphism in this case to be the orbital ridge, zygomatic regions, and frontal bossing area. Based on the 3D template overlay, polyetheretherketone augmentation implants were virtually designed in close consultation with a patient to augment the aforementioned regions. The virtual statistical shape modeling template offered an objective reference, and the possibility to fully involve the patient in the treatment planning.

Semi-automated creation of IFC bridge models from point clouds for maintenance applications

Bridge maintenance activities benefit from digital models, provided in the interoperable IFC format. Such a model, enriched with up-to-date information, is an enabler for a wide range of applications. It opens new perspectives in asset information management. However, the manual creation of a digital replica, representing the actual state of the asset from point cloud data, is time-consuming. Consequently, process automation is of particular interest. This paper proposes a systematic, semi-automatic approach for creating IFC bridge models from point clouds. It introduces new methods for semantic segmentation and 3D shape modeling. A case study demonstrates the feasibility of the process in practice. Compared to other solutions, proposed methods are robust when dealing with incomplete point clouds.

Bridging Formal Shape Models and Deep Learning: A Novel Fusion for Understanding 3D Objects

This article describes a novel fusion of a generative formal model for three-dimensional (3D) shapes with deep learning (DL) methods to understand the geometric structure of 3D objects and the relationships between their components, given a collection of unorganized point cloud measurements. Formal 3D shape models are implemented as shape grammar programs written in Procedural Shape Modeling Language (PSML). Users write PSML programs to describe complex objects, and DL networks estimate the configured free parameters of the program to generate 3D shapes. Users write PSML programs to enforce fundamental rules that define an object class and encode object attributes, including shapes, components, size, position, etc., into a parametric representation of objects. This fusion of the generative model with DL offers artificial intelligence (AI) models an opportunity to better understand the geometric organization of objects in terms of their components and their relationships to other objects. This approach allows human-in-the-loop control over DL estimates by specifying lists of candidate objects, the shape variations that each object can exhibit, and the level of detail or, equivalently, dimension of the latent representation of the shape. The results demonstrate the advantages of the proposed method over competing approaches.

Ethnicity differences in geometric remodelling and myocardial composition in hypertension unveiled by cardiovascular magnetic resonance.

Hypertensive patients of African ancestry (Afr-a) have higher incidences of heart failure and worse clinical outcomes than hypertensive patients of European ancestry (Eu-a), yet the underlying mechanisms remain misunderstood. This study investigated right (RV) and left (LV) ventricular remodelling alongside myocardial tissue derangements between Afr-a and Eu-a hypertensives. 63 Afr-a and 47 Eu-a hypertensives underwent multi-parametric cardiovascular magnetic resonance. Biventricular volumes, mass, function, mass/end-diastolic volume (M/V) ratios, T2 and pre-/post-contrast T1 relaxation times, synthetic extracellular volume, and myocardial fibrosis (MF) were measured. 3D shape modelling was implemented to delineate ventricular geometry. LV and RV mass (indexed to body-surface-area) and M/V ratio were significantly greater in Afr-a than Eu-a hypertensives (67.1 ± 21.7 vs. 58.3 ± 16.7 g/m2, 12.6 ± 3.48 vs. 10.7 ± 2.71 g/m2, 0.79 ± 0.21 vs. 0.70 ± 0.14 g/mL, and 0.16 ± 0.04 vs. 0.13 ± 0.03 g/mL, respectively; P < 0.03). Afr-a patients showed greater basal interventricular septum thickness than Eu-a patients, influencing LV hypertrophy and RV cavity changes. This biventricular remodelling was associated with prolonged T2 relaxation time (47.0 ± 2.2 vs. 45.7 ± 2.2 ms, P = 0.005) and higher prevalence (23% vs. 4%, P = 0.001) and extent of MF [2.3 (0.6-14.3) vs. 1.6 (0.9-2.5) % LV mass, P = 0.008] in Afr-a patients. Multivariable linear regression showed that modifiable cardiovascular risk factors and greater end-diastolic volume, but not ethnicity, were independently associated with greater LV mass. Afr-a hypertensives had distinctive biventricular remodelling, including increased RV mass, septal thickening and myocardial tissue abnormalities compared with Eu-a hypertensives. From this study, modifiable cardiovascular risk factors and ventricular geometry, but not ethnicity, were independently associated with greater LV myocardial mass.

StylePart: image-based shape part manipulation

AbstractDirect part-level manipulation of man-made shapes in an image is desired given its simplicity. However, it is not intuitive given the existing manually created cuboid and cylinder controllers. To tackle this problem, we present StylePart, a framework that enables direct shape manipulation of an image by leveraging generative models of both images and 3D shapes. Our key contribution is a shape-consistent latent mapping function that connects the image generative latent space and the 3D man-made shape attribute latent space. Our method “forwardly maps” the image content to its corresponding 3D shape attributes, where the shape part can be easily manipulated. The attribute codes of the manipulated 3D shape are then “backwardly mapped” to the image latent code to obtain the final manipulated image. By using both forward and backward mapping, an user can edit the image directly without resorting to any 3D workflow. We demonstrate our approach through various manipulation tasks, including part replacement, part resizing, and shape orientation manipulation, and evaluate its effectiveness through extensive ablation studies.

Theoretical and Empirical Analysis of a Fast Algorithm for Extracting Polygons from Signed Distance Bounds

Recently, there has been renewed interest in signed distance bound representations due to their unique properties for 3D shape modelling. This is especially the case for deep learning-based bounds. However, it is beneficial to work with polygons in most computer graphics applications. Thus, in this paper, we introduce and investigate an asymptotically fast method for transforming signed distance bounds into polygon meshes. This is achieved by combining the principles of sphere tracing (or ray marching) with traditional polygonization techniques, such as marching cubes. We provide theoretical and experimental evidence that this approach is of the O(N2logN) computational complexity for a polygonization grid with N3 cells. The algorithm is tested on both a set of primitive shapes and signed distance bounds generated from point clouds by machine learning (and represented as neural networks). Given its speed, implementation simplicity, and portability, we argue that it could prove useful during the modelling stage as well as in shape compression for storage.

One-Shot Neural Face Reenactment via Finding Directions in GAN’s Latent Space

In this paper, we present our framework for neural face/head reenactment whose goal is to transfer the 3D head orientation and expression of a target face to a source face. Previous methods focus on learning embedding networks for identity and head pose/expression disentanglement which proves to be a rather hard task, degrading the quality of the generated images. We take a different approach, bypassing the training of such networks, by using (fine-tuned) pre-trained GANs which have been shown capable of producing high-quality facial images. Because GANs are characterized by weak controllability, the core of our approach is a method to discover which directions in latent GAN space are responsible for controlling head pose and expression variations. We present a simple pipeline to learn such directions with the aid of a 3D shape model which, by construction, inherently captures disentangled directions for head pose, identity, and expression. Moreover, we show that by embedding real images in the GAN latent space, our method can be successfully used for the reenactment of real-world faces. Our method features several favorable properties including using a single source image (one-shot) and enabling cross-person reenactment. Extensive qualitative and quantitative results show that our approach typically produces reenacted faces of notably higher quality than those produced by state-of-the-art methods for the standard benchmarks of VoxCeleb1 & 2.

Fast and Simple Statistical Shape Analysis of Pregnant Women Using Radial Deformation of a Cylindrical Template.

Non-rigid deformation of a template to fit 3D scans of human subjects is widely used to develop statistical models of 3D human shapes and poses. Complex optimization problems must be solved to use these models to parameterize scans of pregnant women, thus limiting their use in antenatal point-of-care tools in low-resource settings. Moreover, these models were developed using datasets that did not contain any 3D scans of pregnant women. In this study, we developed a statistical shape model of the torso of pregnant women at greater than 36 weeks of gestation using fast and simple vertex-based deformation of a cylindrical template constrained along the radial direction. The 3D scans were pre-processed to remove noisy outlier points and segment the torso based on anatomical landmarks. A cylindrical template mesh was then fitted onto the segmented scan of the torso by moving each vertex of in the direction of the radial vector. This process is computationally inexpensive taking only 14.80 seconds to deform a template with 9090 vertices. Principal component analysis (PCA) was performed on the deformed vertex co-ordinates to find the directions of maximum variance. The first 10 principal vectors of our model explained 79.03% of the total variance and reconstructed unseen scans with a mean error of 2.43 cm. We also used the PCA weights of the first 10 principal vectors to accurately predict anthropometric measurements of the pregnant women.

Structural Outlier Detection and Zernike-Canterakis Moments for Molecular Surface Meshes-Fast Implementation in Python.

Object retrieval systems measure the degree of similarity of the shape of 3D models. They search for the elements of the 3D model databases that resemble the query model. In structural bioinformatics, the query model is a protein tertiary/quaternary structure and the objective is to find similarly shaped molecules in the Protein Data Bank. With the ever-growing size of the PDB, a direct atomic coordinate comparison with all its members is impractical. To overcome this problem, the shape of the molecules can be encoded by fixed-length feature vectors. The distance of a protein to the entire PDB can be measured in this low-dimensional domain in linear time. The state-of-the-art approaches utilize Zernike-Canterakis moments for the shape encoding and supply the retrieval process with geometric data of the input structures. The BioZernike descriptors are a standard utility of the PDB since 2020. However, when trying to calculate the ZC moments locally, the issue of the deficiency of libraries readily available for use in custom programs (i.e., without relying on external binaries) is encountered, in particular programs written in Python. Here, a fast and well-documented Python implementation of the Pozo-Koehl algorithm is presented. In contrast to the more popular algorithm by Novotni and Klein, which is based on the voxelized volume, the PK algorithm produces ZC moments directly from the triangular surface meshes of 3D models. In particular, it can accept the molecular surfaces of proteins as its input. In the presented PK-Zernike library, owing to Numba's just-in-time compilation, a mesh with 50,000 facets is processed by a single thread in a second at the moment order 20. Since this is the first time the PK algorithm is used in structural bioinformatics, it is employed in a novel, simple, but efficient protein structure retrieval pipeline. The elimination of the outlying chain fragments via a fast PCA-based subroutine improves the discrimination ability, allowing for this pipeline to achieve an 0.961 area under the ROC curve in the BioZernike validation suite (0.997 for the assemblies). The correlation between the results of the proposed approach and of the 3D Surfer program attains values up to 0.99.

A Temporal Learning Approach to Inpainting Endoscopic Specularities and Its Effect on Image Correspondence

Video streams are utilised to guide minimally-invasive surgery and diagnosis in a wide range of procedures, and many computer-assisted techniques have been developed to automatically analyse them. These approaches can provide additional information to the surgeon such as lesion detection, instrument navigation, or anatomy 3D shape modelling. However, the necessary image features to recognise these patterns are not always reliably detected due to the presence of irregular light patterns such as specular highlight reflections. In this paper, we aim at removing specular highlights from endoscopic videos using machine learning. We propose using a temporal generative adversarial network (GAN) to inpaint the hidden anatomy under specularities, inferring its appearance spatially and from neighbouring frames, where they are not present in the same location. This is achieved using in-vivo data from gastric endoscopy (Hyper Kvasir) in a fully unsupervised manner that relies on the automatic detection of specular highlights. System evaluations show significant improvements to other methods through direct comparison and ablation studies that depict the importance of the network’s temporal and transfer learning components. The generalisability of our system to different surgical setups and procedures was also evaluated qualitatively on in-vivo data of gastric endoscopy and ex-vivo porcine data (SERV-CT, SCARED). We also assess the effect of our method in comparison to other methods on computer vision tasks that underpin 3D reconstruction and camera motion estimation, namely stereo disparity, optical flow, and sparse point feature matching. These are evaluated quantitatively and qualitatively and results show a positive effect of our specular inpainting method on these tasks in a novel comprehensive analysis. Our code and dataset are made available at https://github.com/endomapper/Endo-STTN.

Precise 3D profile determination of high aspect ratio hole patterns by transmission small-angle X-ray scattering

The etching process of high aspect ratio (HAR) hole patterns on a wafer surface is a key step for fabricating new-generation semiconductor memory devices with vertically stacked structures. As the stacking number of these memory devices increases, it is getting more challenging to maintain the ideal etching profile of HAR holes. Therefore, detailed profile evaluation of these HAR holes is increasingly important. In this study, we have measured 4.2 μm deep holes by transmission small-angle X-ray scattering (T-SAXS) to determine the precise three-dimensional (3D) hole profile. By applying an improved 3D shape model for a hole, we successfully determined a hole profile whose cross-section parallel to the sample surface changes from elliptical to rectangular along its depth. This 3D profile measurement demonstrated that T-SAXS has sufficient sensitivity to evaluate a cross-sectional shape change along the depth of HAR holes.

Three-dimensional assessment of interfractional cervical and uterine motions using daily magnetic resonance images to determine margins and timing of replanning.

This study was conducted to determine the margins and timing of replanning by assessing the daily interfractional cervical and uterine motions using magnetic resonance (MR) images. Eleven patients with cervical cancer, who underwent intensity-modulated radiotherapy (IMRT) in 23-25 fractions, were considered in this study. The daily and reference MR images were converted into three-dimensional (3D) shape models. Patient-specific anisotropic margins were calculated from the proximal 95% of vertices located outside the surface of the reference model. Population-based margins were defined as the 90th percentile values of the patient-specific margins. The expanded volume of interest (expVOI) for the cervix and uterus was generated by expanding the reference model based on the population-based margin to calculate the coverage for daily deformable mesh models. For comparison, expVOIconv was generated using conventional margins: right (R), left (L), anterior (A), posterior (P), superior (S), and inferior (I) were (5, 5, 15, 15, 10, 10) and (10, 10, 20, 20, 15, 15) mm for the cervix and uterus, respectively. Subsequently, a replanning scenario was developed based on the cervical volume change. ExpVOIini and expVOIreplan were generated before and after replanning, respectively. Population-based margins were (R, L, A, P, S, I) of (7, 7, 11, 6, 11, 8) and (14, 13, 27, 19, 15, 21) mm for the cervix and uterus, respectively. The timing of replanning was found to be the 16th fraction, and the volume of expVOIreplan decreased by >30% compared to that of expVOIini . However, margins cannot be reduced to ensure equivalent coverage after replanning. We determined the margins and timing of replanning through detailed daily analysis. The margins of the cervix were smaller than conventional margins in some directions, while the margins of the uterus were larger in almost all directions. A margin equivalent to that at the initial planning was required for replanning.

A 3‐D Shape Reconstruction Strategy for Small Solar System Bodies With Single Flyby Spaceborne Radar

AbstractThis article presents a novel 3‐D shape reconstruction strategy for small solar system bodies using spaceborne radar which is applied in flyby trajectories to characterize multiple targets in one mission to save time and cost. In this study, we establish a measurement geometry for rapidly spinning small solar system bodies that combines the motion properties of synthetic aperture radar and circular synthetic aperture radar techniques. We also propose a modified imaging algorithm, which is processed using Cardano's formula and the modified polar format algorithm, to improve the imaging quality. Finally, we present simulated results of point‐scattering and 3D shape models. Our analysis shows that the shape reconstruction errors of the algorithm are less than 1.5 m as indicated by the standard deviation.

3D patellar shape is associated with radiological and clinical signs of patellofemoral osteoarthritis

To examine the association between 3D patellar shape and 1) isolated magnetic resonance imaging (MRI)-based patellofemoral osteoarthritis (PFOA), 2) the morphological features of PFOA, and 3) the clinical symptoms of PFOA. MRIdata from 66 women with isolated MRI-based PFOA and 66 age- and BMI-matched healthy women were selected from a cohort study. The patellae were manually segmented from MRI scans and used to create a 3D statistical shape model (SSM) of the patella. Structural abnormalities were semi-standardized scored on MRI using MRI osteoarthritis knee score (MOAKS). Regression analyses were applied to determine the associations between the shape parameters retrieved from the SSM, group status, clinical symptoms, and structural abnormalities. Four shape variants showed a statistically significant (<0.05) association with the group status. The mode responsible for most of the shape variations showed participants with PFOA possess a relatively thicker dorsal bump on the articular part of the patella, compared to patellae of control participants. Three of these variants showed an association with the presence of osteophytes and cartilage loss on the patella. Multiple associations were found between patellar shape and the clinical symptoms of PFOA. Patellar shape is associated with the prevalence of MRI-based PFOA in women. Some shape variants were also associated with clinical symptoms. Interestingly, one particular shape variant associated with the presence of MRI-based PFOA was earlier shown to be associated with structural abnormalities associated with OA in a population aged under 40. This may suggest that patellar shape may be an early detectable risk factor for PFOA.

Soft body reconstruction of a reptile fossil by the nondestructive elemental mapping with a newly designed XRF

AbstractVertebrate fossils usually consist only of mineralized skeletons and rarely preserve soft body parts. However, the buried soft body releases the degraded material into the surrounding sediments. The degraded material can be preserved as invisible signals of some chemicals or authigenic minerals, which can be revealed by the elemental distribution of the fossils' surface and their surrounding rock. As a method for surface elemental mapping, X‐ray fluorescence (XRF) measurement requires the samples to be ground and polished as a mirror‐like plane that keeps a stable distance of X‐ray source‐sample‐detector. This destructive pretreatment is undesirable for precious fossils. To rectify this problem, a nondestructive XRF for element mapping of the fossils' 3D surface is newly invented. The new equipment can follow the 3D shape model to keep the distance for every XRF scanned point. We performed the element mapping of Ca and Fe on the surface of a marine sauropterygian fossil (Keichousaurus hui) and its surrounding sediments, and found that the soft tissue around the thoracic cage is thicker than the abdomen. The abnormal elemental distribution on the area of the angular and surangular of the skull can be explained as unprofessionally prepared. In addition, this research found that elemental mapping revealed the invisible signals of the paleoenvironment around fossils and information for the reconstruction of soft‐tissue anatomy. We also give an example that based on the wide application of the “Internet of Things” and “Industrial Network,” the design and development of specialized equipment for paleontologists' requirements are going simple.

Fast 3D Face Reconstruction from a Single Image Using Different Deep Learning Approaches for Facial Palsy Patients.

The 3D reconstruction of an accurate face model is essential for delivering reliable feedback for clinical decision support. Medical imaging and specific depth sensors are accurate but not suitable for an easy-to-use and portable tool. The recent development of deep learning (DL) models opens new challenges for 3D shape reconstruction from a single image. However, the 3D face shape reconstruction of facial palsy patients is still a challenge, and this has not been investigated. The contribution of the present study is to apply these state-of-the-art methods to reconstruct the 3D face shape models of facial palsy patients in natural and mimic postures from one single image. Three different methods (3D Basel Morphable model and two 3D Deep Pre-trained models) were applied to the dataset of two healthy subjects and two facial palsy patients. The reconstructed outcomes were compared to the 3D shapes reconstructed using Kinect-driven and MRI-based information. As a result, the best mean error of the reconstructed face according to the Kinect-driven reconstructed shape is 1.5±1.1&amp;nbsp;mm. The best error range is 1.9±1.4&amp;nbsp;mm when compared to the MRI-based shapes. Before using the procedure to reconstruct the 3D faces of patients with facial palsy or other facial disorders, several ideas for increasing the accuracy of the reconstruction can be discussed based on the results. This present study opens new avenues for the fast reconstruction of the 3D face shapes of facial palsy patients from a single image. As perspectives, the best DL method will be implemented into our computer-aided decision support system for facial disorders.

Synthetic OCT Data Generation to Enhance the Performance of Diagnostic Models for Neurodegenerative Diseases.

PurposeOptical coherence tomography (OCT) has recently emerged as a source for powerful biomarkers in neurodegenerative diseases such as multiple sclerosis (MS) and neuromyelitis optica (NMO). The application of machine learning techniques to the analysis of OCT data has enabled automatic extraction of information with potential to aid the timely diagnosis of neurodegenerative diseases. These algorithms require large amounts of labeled data, but few such OCT data sets are available now.MethodsTo address this challenge, here we propose a synthetic data generation method yielding a tailored augmentation of three-dimensional (3D) OCT data and preserving differences between control and disease data. A 3D active shape model is used to produce synthetic retinal layer boundaries, simulating data from healthy controls (HCs) as well as from patients with MS or NMO.ResultsTo evaluate the generated data, retinal thickness maps are extracted and evaluated under a broad range of quality metrics. The results show that the proposed model can generate realistic-appearing synthetic maps. Quantitatively, the image histograms of the synthetic thickness maps agree with the real thickness maps, and the cross-correlations between synthetic and real maps are also high. Finally, we use the generated data as an augmentation technique to train stronger diagnostic models than those using only the real data.ConclusionsThis approach provides valuable data augmentation, which can help overcome key bottlenecks of limited data.Translational RelevanceBy addressing the challenge posed by limited data, the proposed method helps apply machine learning methods to diagnose neurodegenerative diseases from retinal imaging.