Compressed Representation of 3D Models for 3D Printing

Abstract

In this work we tackle the issue of large file sizes in disk of volume data representations of 3D models when used for 3D printing purposes. In the pipeline of 3D printing software, the representation, in disk, of the model go from the initial representation, which normally is a polygonal mesh, to volume data and later into instructions for the printer. Each of these stages has a cost in terms of size of disk and memory usage, and with large models with high resolutions, the size of the representations can become unfeasible to save and the memory usage may be too large. For that reason, a better encoding of the representation of volume data is necessary, while maintaining the efficiency of the processing of slices and not impacting or reducing the memory usage of the creating of such encoding. We describe several methods with different configurations utilised to reduce the size of the representation in disk, such as Run Length encoding and Modulus Encoding, while also compressing them with known lossless compression algorithms such as LZMA, used by 7Zip, and DEFLATE, used by GZip. We utilise several models to show that our results greatly reduce the size of the representation without impacting any other stage of the pipeline, and therefore providing a better alternative to current market techniques.