Session: CIE-07-01: CAPPD General

Paper Number: 116501

116501 - Parameter Extraction From Images Using Multilabel Supervised Learning 

In this work, we propose an approach to predict multiple design parameters of products using 2D images and supervised
learning techniques. Fully parametric 2D or 3D vector representations have a high degree of flexibility since they are generally
capable of morphing between all designs in the design space. A reverse design method is applied to parametrically designed
products to extract design knowledge from existing designs and apply it to future designs. The goal of this research is to develop
an accurate and efficient predictive model for multiple design parameters. The paper introduces several novel aspects, including
learning parametric patterns from a 2D model dataset and using them to predict design parameters for new engineering sys-
tems. Additionally, the paper proposes a finetuning multilabel approach for predicting multiple design parameters of different
types simultaneously, including binary and continuous parameters. Results demonstrate that the proposed approach achieves
high prediction accuracy and outperforms existing methods for predicting multiple product design parameters. The findings sug-
gest that the proposed approach can be used as a reliable tool for predicting multiple design parameters of products in practi-
cal engineering design scenarios from only 2D images, potentially accelerating the design process and enabling designers to
explore a wider range of design possibilities.

Presenting Author: Jessica Ezemba Carnegie Mellon University

Presenting Author Biography: Jessica is a graduate student in Mechanical Engineering. Her research focus included generative design using AI/ML techniques

Authors:

Jessica Ezemba Carnegie Mellon University
James Cunningham Carnegie Mellon University
Conrad Tucker Carnegie Mellon University