Abstract
Computer-Aided Design (CAD) and Computer-Aided Manufacturing (CAM) systems have revolutionized modern mechanical design and production processes. A critical step in streamlining manufacturing is feature recognition, which identifies geometric and functional attributes of components to automate downstream processes. Traditional feature recognition techniques rely on rule based and geometric algorithms, which are limited in handling complex, freeform, or feature-interacting models. Deep learning (DL), a subset of artificial intelligence, offers remarkable potential in automating feature recognition by learning patterns from large datasets without explicit programming. This review paper discusses the current advancements in deep learning-based automated feature recognition in CAD/CAM, highlighting techniques such as convolutional neural networks (CNNs), graph neural networks (GNNs), and hybrid approaches. The challenges, datasets, evaluation metrics, and future research directions are also analyzed. The paper aims to provide researchers and engineers with a comprehensive understanding of deep learning applications for feature recognition in modern CAD/CAM systems.
Keywords: CAD/CAM, Deep Learning, Feature Recognition, Convolutional Neural Networks, Graph Neural Networks, Automated Manufacturing
Full Issue
| View or download the full issue | PDF 133-148 |