CAD programs, structured as parametric sequences of commands that compile into precise 3D geometries, are fundamental to accurate and efficient engineering design processes. Generating these programs from nonparametric data such as point clouds and meshes remains a crucial yet challenging task, typically requiring extensive manual intervention. Current deep generative models aimed at automating CAD generation are significantly limited by imbalanced and insufficiently large datasets, particularly those lacking representation for complex CAD programs. To address this, we introduce GenCAD-3D, a multimodal generative framework utilizing contrastive learning for aligning latent embeddings between CAD and geometric encoders, combined with latent diffusion models for CAD sequence generation and retrieval. Additionally, we present SynthBal, a synthetic data augmentation strategy specifically designed to balance and expand datasets, notably enhancing representation of complex CAD geometries. Our experiments show that SynthBal significantly boosts reconstruction accuracy, reduces the generation of invalid CAD models, and markedly improves performance on high-complexity geometries, surpassing existing benchmarks. These advancements hold substantial implications for streamlining reverse engineering and enhancing automation in engineering design. We will publicly release our datasets and code, including a set of 51 3D-printed and laser-scanned parts on our project site.

Efficient creation of accurate and editable 3D CAD models is critical in engineering design, significantly impacting cost and time-to-market in product innovation. Current manual workflows remain highly time-consuming and demand extensive user expertise. While recent developments in AI-driven CAD generation show promise, existing models are limited by incomplete representations of CAD operations, inability to generalize to real-world images, and low output accuracy. This paper introduces CAD-Coder, an open-source Vision-Language Model (VLM) explicitly fine-tuned to generate editable CAD code (CadQuery Python) directly from visual input. Leveraging a novel dataset that we created--GenCAD-Code, consisting of over 163k CAD-model image and code pairs--CAD-Coder outperforms state-of-the-art VLM baselines such as GPT-4.5 and Qwen2.5-VL-72B, achieving a 100% valid syntax rate and the highest accuracy in 3D solid similarity. Notably, our VLM demonstrates some signs of generalizability, successfully generating CAD code from real-world images and executing CAD operations unseen during fine-tuning. The performance and adaptability of CAD-Coder highlights the potential of VLMs fine-tuned on code to streamline CAD workflows for engineers and designers. CAD-Coder is publicly available at: https://github.com/anniedoris/CAD-Coder.

Computer-aided design (CAD) tools empower designers to design and modify 3D models through a series of CAD operations, commonly referred to as a CAD sequence. In scenarios where digital CAD files are not accessible, reverse engineering (RE) has been used to reconstruct 3D CAD models. Recent advances have seen the rise of data-driven approaches for RE, with a primary focus on converting 3D data, such as point clouds, into 3D models in boundary representation (B-rep) format. However, obtaining 3D data poses significant challenges, and B-rep models do not reveal knowledge about the 3D modeling process of designs. To this end, our research introduces a novel data-driven approach with an Image2CADSeq neural network model. This model aims to reverse engineer CAD models by processing images as input and generating CAD sequences. These sequences can then be translated into B-rep models using a solid modeling kernel. Unlike B-rep models, CAD sequences offer enhanced flexibility to modify individual steps of model creation, providing a deeper understanding of the construction process of CAD models. To quantitatively and rigorously evaluate the predictive performance of the Image2CADSeq model, we have developed a multi-level evaluation framework for model assessment. The model was trained on a specially synthesized dataset, and various network architectures were explored to optimize the performance. The experimental and validation results show great potential for the model in generating CAD sequences from 2D image data.

The creation of manufacturable and editable 3D shapes through Computer-Aided Design (CAD) remains a highly manual and time-consuming task, hampered by the complex topology of boundary representations of 3D solids and unintuitive design tools. This paper introduces GenCAD, a generative model that employs autoregressive transformers and latent diffusion models to transform image inputs into parametric CAD command sequences, resulting in editable 3D shape representations. GenCAD integrates an autoregressive transformer-based architecture with a contrastive learning framework, enhancing the generation of CAD programs from input images and providing a representation learning framework for multiple data modalities relevant to engineering designs. Extensive evaluations demonstrate that GenCAD significantly outperforms existing state-of-the-art methods in terms of the precision and modifiability of generated 3D shapes. Notably, GenCAD shows a marked improvement in the accuracy of 3D shape generation for long sequences, supporting its application in complex design tasks. Additionally, the contrastive embedding feature of GenCAD facilitates the retrieval of CAD models using image queries from databases which is a critical challenge within the CAD community. While most work in the 3D shape generation literature focuses on representations like meshes, voxels, or point clouds, practical engineering applications demand modifiability and the ability for multi-modal conditional generation. Our results provide a significant step forward in this direction, highlighting the potential of generative models to expedite the entire design-to-production pipeline and seamlessly integrate different design modalities.

Computer-Aided Design (CAD) has been around since the 1950s. The first graphical CAD program, called Sketchpad, came out of MIT (designworldonline.com). Since then, CAD has become essential to designing and manufacturing hardware products. Today, there are multiple types of CAD. This article focuses on mechanical CAD, used for mechanical engineering. Digging into the history of computer graphics reveals some interesting connections between the most ambitious and notorious engineers. Ivan Sutherland, who received the Turing Award for Sketchpad in 1988, had Edwin Catmull as a student.

The notion of using computers to improve cancer diagnostics is decades old, and computer-aided detection (CAD) systems are commonplace in mammography clinics, yet CAD programs have not improved performance in clinical practice. The issue, Dr Lehman believes, is current CAD programs were trained to identify things human radiologists can see, whereas AI learns to spot cancers based on the actual results of thousands of mammograms. This has the potential to "exceed human capacity to identify subtle cues that the human eye and brain aren't able to perceive," she added.

CHICAGO (Reuters) - A Google artificial intelligence system proved as good as expert radiologists at detecting which women had breast cancer based on screening mammograms and showed promise at reducing errors, researchers in the United States and Britain reported. The study, published in the journal Nature on Wednesday, is the latest to show that artificial intelligence (AI) has the potential to improve the accuracy of screening for breast cancer, which affects one in eight women globally. Radiologists miss about 20% of breast cancers in mammograms, the American Cancer Society says, and half of all women who get the screenings over a 10-year period have a false positive result. The findings of the study, developed with Alphabet Inc's (GOOGL.O) DeepMind AI unit, which merged with Google Health in September, represent a major advance in the potential for the early detection of breast cancer, Mozziyar Etemadi, one of its co-authors from Northwestern Medicine in Chicago, said. The team, which included researchers at Imperial College London and Britain's National Health Service, trained the system to identify breast cancers on tens of thousands of mammograms.

In the United States, only one radiologist reads the results and the tests are done every one to two years. In Britain, the tests are done every three years, and each is read by two radiologists. When they disagree, a third is consulted.'SUBTLE CUES'In a separate test, the group pitted the AI system against six radiologists and found it outperformed them at accurately detecting breast cancers.Connie Lehman, chief of the breast imaging department at Harvard's Massachusetts General Hospital, said the results are in line with findings from several groups using AI to improve cancer detection in mammograms, including her own work.The notion of using computers to improve cancer diagnostics is decades old, and computer-aided detection (CAD) systems are commonplace in mammography clinics, yet CAD programs have not improved performance in clinical practice.The issue, Lehman said, is that current CAD programs were trained to identify things human radiologists can see, whereas with AI, computers learn to spot cancers based on the actual results of thousands of mammograms.This has the potential to "exceed human capacity to identify subtle cues that the human eye and brain aren't able to perceive," Lehman added.Although computers have not been "super helpful" so far, "what we've shown at least in tens of thousands of mammograms is the tool can actually make a very well-informed decision," Etemadi said.The study has some limitations. Most of the tests were done using the same type of imaging equipment, and the U.S. group contained a lot of patients with confirmed breast cancers.Crucially, the team has yet to show the tool improves patient care, said Dr Lisa Watanabe, chief medical officer of CureMetrix, whose AI mammogram program won U.S. approval last year."AI

CHICAGO – A Google artificial intelligence system proved as good as expert radiologists at predicting which women would develop breast cancer based on screening mammograms and showed promise at reducing errors, researchers in the United States and Britain reported. The study, published in the journal Nature on Wednesday, is the latest to show that artificial intelligence (AI) has the potential to improve the accuracy of screening for breast cancer, which affects one in eight women globally. Radiologists miss about 20 percent of breast cancers in mammograms, the American Cancer Society says, and half of all women who get the screenings over a 10-year period have a false positive result. The findings of the study, developed with Alphabet's DeepMind AI unit, which merged with Google Health in September, represent a major advance in the potential for the early detection of breast cancer, said Mozziyar Etemadi, one of its co-authors from Northwestern Medicine in Chicago. The team, which included researchers at Imperial College London and Britain's National Health Service, trained the system to identify breast cancers on tens of thousands of mammograms.

Recently, multiple formulations of vision problems as probabilistic inversions of generative models based on computer graphics have been proposed. However, applications to 3D perception from natural images have focused on low-dimensional latent scenes, due to challenges in both modeling and inference. Accounting for the enormous variability in 3D object shape and 2D appearance via realistic generative models seems intractable, as does inverting even simple versions of the many-to-many computations that link 3D scenes to 2D images. This paper proposes and evaluates an approach that addresses key aspects of both these challenges. We show that it is possible to solve challenging, real-world 3D vision problems by approximate inference in generative models for images based on rendering the outputs of probabilistic CAD (PCAD) programs. Our PCAD object geometry priors generate deformable 3D meshes corresponding to plausible objects and apply affine transformations to place them in a scene. Image likelihoods are based on similarity in a feature space based on standard mid-level image representations from the vision literature. Our inference algorithm integrates single-site and locally blocked Metropolis-Hastings proposals, Hamiltonian Monte Carlo and discriminative data-driven proposals learned from training data generated from our models. We apply this approach to 3D human pose estimation and object shape reconstruction from single images, achieving quantitative and qualitative performance improvements over state-of-the-art baselines.