BIKED++: A Multimodal Dataset of 1.4 Million Bicycle Image and Parametric CAD Designs
This dataset addresses the need for large-scale multimodal data to train cross-modal predictive models in design and engineering, though it is incremental as it builds on the existing BIKED dataset family.
The paper introduces BIKED++, a public dataset of 1.4 million procedurally-generated bicycle designs with parametric, JSON, and image representations, created using a rendering engine based on BikeCAD software, and demonstrates its use by training a model to estimate CLIP embeddings from parametric designs for cross-modal similarity tasks.
This paper introduces a public dataset of 1.4 million procedurally-generated bicycle designs represented parametrically, as JSON files, and as rasterized images. The dataset is created through the use of a rendering engine which harnesses the BikeCAD software to generate vector graphics from parametric designs. This rendering engine is discussed in the paper and also released publicly alongside the dataset. Though this dataset has numerous applications, a principal motivation is the need to train cross-modal predictive models between parametric and image-based design representations. For example, we demonstrate that a predictive model can be trained to accurately estimate Contrastive Language-Image Pretraining (CLIP) embeddings from a parametric representation directly. This allows similarity relations to be established between parametric bicycle designs and text strings or reference images. Trained predictive models are also made public. The dataset joins the BIKED dataset family which includes thousands of mixed-representation human-designed bicycle models and several datasets quantifying design performance. The code and dataset can be found at: https://github.com/Lyleregenwetter/BIKED_multimodal/tree/main