Skip to content
On this page


Optical Chemical Structure Recognition (OCSR) is a technology that automatically extracts chemical structures from images or scanned documents, enabling efficient digitization and analysis of chemical data.

We packaged CM with DECIMER, an open-source software platform for automated optical chemical structure identification, segmentation and recognition in scientific publications. It is a deep learning-based system that uses a transformer-based algorithm to convert the bitmap of a chemical structure depiction into a computer-readable format. DECIMER is trained on a massive dataset of chemical structures, and it can be used to identify and recognize chemical structures in scanned scientific literature.