SDLG<!-- --> | Studio Coat<!-- -->

← Back


Research | On | Github



SDLG

SDLG (Scalable Deep Learning Graphics) is a proof of concept for using an LSTM network to generate text in an SVG style to create vector logos.


thumbnail

120 svg files have been converted to text was utilized as the data source. The svg files that were picked are path-based svg files. Other elements in an svg file that contribute to the formation of a logo include polygons, circles, rectangles, and so on. As a result, I've decided to solely use pathways because the network will be easier to train.

thumbnail
thumbnail
thumbnail
thumbnail
© 2022 Studio Coat