ShapeEmbed: a self-supervised learning framework for 2D contour quantification | Read Paper on Bytez